Output 3: Grass and legume genotypes with superior adaptation to edaphic and climatic constraints are developed Activity 3.1 Genotypes of Brachiaria, Panicum, and Arachis with adaptation to edaphic and climatic factors Highlights • Found that two sexual Brachiaria hybrids (SX 2349 and SX 497) had greater level of Al resistance than sexual parent (BRUZ/44-02). • Showed that the superior performance of the Brachiaria hybrid (FM9503-S046-024) after establishment was associated with its ability to acquire greater amounts of nutrients from low fertility soil. • Showed that B. humidicola suppresses nitrification and nitrous oxide emission by inhibiting the activity of ammonium oxidizing bacteria in the soil. • Using phosphorus isotope exchange kinetics technique, found that Brachiaria decumbens CIAT 606 acquires P from less available P forms in an oxisol. • Field evaluation over 3 years showed that Arachis pintoi CIAT 22159 may be better than the commercial cultivar (CIAT 17434) in terms of persistence with no P fertilizer input. Progress towards achieving milestones • New Brachiaria sexual hybrids with Al resistance identified We were successful in identifying 2 sexual hybrids (SX 2349 and SX 497) with greater level of Al resistance than that of the sexual parent, BRUZ/44-02. These sexual hybrids are currently being used in the on-going breeding program to combine spittlebug resistance with high level of Al resistance. • Brachiaria hybrids with superior performance under low soil fertility identified We selected a Brachiaria hybrid, (FM9503-S046-024) for its excellent adaptation to low fertility soils, which was associated with its ability to acquire greater amounts of nutrients from the soil as compared to other hybrids. Using phosphorus isotope exchange kinetics technique in collaboration with scientists from ETH, Zurich Switzerland, we showed that Brachiaria decumbens CIAT 606 acquires P from less available P forms from low P oxisol. Further research work is needed to evaluate the ability of most promising Brachiaria hybrids for this unique ability to acquire less available P forms from low fertility oxisols. • Ability to suppress nitrification and emission of nitrous oxide by Brachiaria humidicola quantified Collaborative research with JIRCAS scientists from Japan showed that B. humidicola suppresses nitrification and nitrous oxide emission by inhibiting the activity of ammonium oxidizing bacteria in the soil. Further research work is in progress to determine genotypic variation in this ability of B. humidicola. • Field experiment with promising accessions of Arachis pintoi established in the Llanos We identified Arachis pintoi CIAT 22159 as a promising accession for targeting to infertile acid soils. This accession along with a few other accessions (CIAT 18744, 18748, 22160) and commercial cultivar (CIAT 17434) are being evaluated in relatively better soils of the Llanos of Colombia (Piedmont region) for their suitability as cover or forage legumes. Low supply of nutrients and aluminum (Al) toxicity are major limitations to tropical forage production on acid soils. Previous research indicated that tropical grasses and legumes that are adapted to infertile acid soils have root and shoot attributes that are linked to strategies to acquire nutrients in a low pH and high Al environment. Identification of those key plant attributes is fundamental to develop more efficient screening procedures for germplasm evaluation and/or improvement. 3.1.1 Development of improved tetraploid, sexual Brachiaria hybrid breeding population for resistance to edaphic factors and general environmental adaptation Contributors: I. M. Rao, J. W. Miles, P. Wenzl, R. García and J. Ricaurte (CIAT) The first criterion for selection in the tetraploid, sexual Brachiaria breeding population is performance in field trials on acid, Al-toxic soils at CIAT-Quilichao and the Matazul farm in Puerto López (Llanos Orientales). Over 4,300 progeny plants were established during 2001 in duplicate, space-planted field trials. Plants are being culled on visual assessment of vigor and freedom of obvious deficiency symptoms as well as other attributes. A manageable set of fewer than 1000 clones will be identified by 01 December 2001 and propagated for in vitro assessment of Al tolerance (and other attributes) during first semester of 2002, to identify a small set of parental sexual clones to be intercrossed in isolation to produce the subsequent cycle population. 3.1.2 Studies on mechanisms of acid soil adaptation in Brachiaria cultivars and development of screening methods 3.1.2.1 Identification of Al-resistant Brachiaria hybrids Contributors: I.M. Rao, J. W. Miles, P. Wenzl, R. García and J. Ricaurte (CIAT) Rationale Last year, we implemented screening procedure to identify Al-resistant Brachiaria hybrids that were preselected for spittlebug resistance. As part of the restricted core project funded by BMZ-GTZ of Germany, this year we used this screening method to identify most promising sexual Brachiaria hybrids that combine Al resistance with spittlebug resistance. Materials and Methods A total of 46 genotypes including 3 parents (B. decumbens CIAT 606, B. brizantha CIAT 6294 and B. ruziziensis 44-02) were selected for evaluation of Al resistance. Among the 43 hybrids selected for screening, 41 were sexuals and two hybrids, CIAT 36061 and CIAT 36062 were apomicts. All the hybrids except CIAT 36061 were highly resistant to spittlebug infestation (C. Cardona, personal communication). Stem-cuttings were rooted in a CaCl2 (200µM) solution, selected for uniformity and transferred to a solution containing 200µM CaCl2 (pH 4.2) and exposed to 2 levels of AlCl3 (0 and 200µM). The solution was replaced every third day, and total root length and root biomass were measured after 21 days of Al treatment. Results and Discussion As observed before, results on total root length indicate that the parent B. decumbens CIAT 606 is outstanding in its level of Al resistance (Figure 29). Among the 43 hybrids tested, 2 apomictic hybrids (CIAT 36061, CIAT 36062) and 2 sexual hybrids (SX 2349 and SX 497) showed moderate level of Al 110 resistance (Figure 29). The level of Al resistance of these two sexuals was markedly superior to that of the sexual parent, BRUZ/44-02. Among the hybrids, CIAT 36061 showed greater level of Al resistance than that of the other hybrids. The relationship between root length with no Al and root length with high Al showed that the extent of total root length of the apomictic Brachiaria hybrid cv. Mulato (CIAT 36061) was similar to that of the outstanding parent B. decumbens CIAT 606 with no Al in solution (Figure 30). But in the presence of high level of Al in solution this hybrid did not perform as well as CIAT 606 but markedly superior to the rest of the hybrids. The hybrids that were identified in the upper box of the right hand side (Figure 30) are being utilized in the on-going breeding program. The root apex (root cap, meristem, and elongation zone) accumulates more Al and attracts greater physical damage than the mature root tissues. Scanning electron microscopic observations revealed differences in Al toxicity effects on root tips of two parents and one hybrid (Photo 13 ). The extent of deformation caused by Al toxicity was minimum with B. decumbens and was marked with B. ruziziensis. The hybrid, CIAT 36061 showed relatively less deformation of root tips compared with B. ruziziensis. Conclusions We have identified 2 sexual hybrids (SX 2349 and SX 497) with greater level of Al resistance than that of the sexual parent, BRUZ/44-02. We are also in the process of evaluating a hybrid population of B. decumbens x B. ruziziensis. This will enable us to develop molecular markers for Al resistance in Brachiaria. 12 Low Al; LSD0.05 = 2.846 High Al; LSD0.05 = 2.167 10 8 6 4 2 0 Genotype Figure 29. Screening for Al resistance among 46 genotypes of Brachiaria. Total root length was measured after exposure to 0 or 200 µM AlCl3 with 200 µM CaCl2 (pH 4.2) for 21 days. 111 Root length (m plant-1) C C II A A TT CI 60 A 3 T 6 6 06 S C 36 2 X IA 0 9 T 61 S 9N 62 X O 9 S 9 /2 4 X 9 9 N 3 9 O 4 9 S NO 4 X99 9 1 7 S 3 X N 9 O 70 S 9 7X NO 3 9 1 SX 9N 8 9 O 3 9 5 SX N 5 9 O 7 S 9N 2 4 X9 O 6 S 9N 2 2 1 X O 3 9 5 S 9N 2 4 X 8 9 O 29 2 SX N 16 9 O 39 0 S N 2 O 5 X S 9 1 1 4 X 9 9 N 1 O 4 S 9 5 NO 8 X 2 9 2 3 SX 9N 29 8 S 9 O 0 X N 9 O 27 9 1 5 SX N 59 O 1 S 9 3 X N 3 9 O 7 S 9 7 2 0 X NO 173 S 99 X N 2 0 O 2 S 0 0 2 0 X N 9 O 6 6 2 3 S 9N X O 6 S 99 0 2 6 X 1 9 N 9 O 62 B NO 246 S R X U 2 9 Z 1 9 S 4 1 4 5 X N 9 O - 0 1 2 S 9NO 8 X S 99 3 3 3 X N 5O 64 S 99 X N O 7 S 99N 1 1 1 X9 O 62 S 9 2 X N 1 9 O 616 S 9N 1 X9 O 34 S 9 5 X N 3 9 O 4 8 9 2 8 SX N 9 O 92 S 9 1 7 X N9 O 2 6 S 9 0 X N 2 9 O 03 9N 1 0 S O 805 S X9 2 X 9 89 N 57 S 9N OO X 29 S 99 3 X N 6 9 O 90 S X 9 1 9 NO 6 9 4 NO 2 3 2 6 341 10 Mean CIAT 606 8 6 CIAT 36062 CIAT 36061 4 SX99NO 2349 CIAT 6294 SX99NO 497 SX99NO 1370 SX99NO 835 SX99NO 574 SX99NO 731 2 SX99NO 2621 SX99NO 15S1X399NO 2822 SX99NO 1630 SX99NO 2354 BRUZ 44-02 SX99NO 2663 SX99NO 2514 SX99NO 2173 SX99NO 2162 Mean SX99NO1260 SX99NO 1345 SX99NO 2927 SX99NO 1616 0 SX99NO 2341 0 2 4 6 8 10 Root length without Al (m plant-1) Figure 30. Identification of Al resistant Brachiaria hybrids. Hybrids that were superior in root length with no or high Al in solution were identified in the upper box of the right hand side. Total root length was measured after exposure to 0 or 200 µM AlCl3 with 200 µM CaCl2 (pH 4.2) for 21 days. 0 µM Al 200 µM Al B. decumbens CIAT 606 B. ruziziensis 44-02 200 µM Al Brachiaria hybrid CIAT 36061 Photo 13. Scanning electron micrographs of root tips of two parents (B. decumbens and B. ruziziensis) and one hybrid (CIAT 36061) of Brachiaria exposed to either 0 or 200 µM of AlCl3 for 21 days. 112 Root length with high Al (m plant-1) 3.1.2.2 Identification of genetic recombinants of Brachiaria with tolerance to low nutrient supply Contributors: I.M. Rao, J. W. Miles, C. Plazas, J. Racaurte and R. García (CIAT) Rationale A field study was conducted at Matazul Farm in the Llanos of Colombia. The main objective was to identify genetic recombinants of Brachiaria with tolerance to low nutrient supply and evaluate plant attributes that contribute to superior adaptation. Materials and Methods A field trial was established on a sandy loam oxisol at Matazul farm in the Llanos of Colombia in July, 1999. The trial comprises 12 entries, including six natural accessions (four parents) and six genetic recombinants of Brachiaria. The trial was planted as a randomized block in split-plot arrangement with two levels of initial fertilizer application (low: kg/ha of 20P, 20K, 33Ca, 14 Mg, 10S; and high: 80N, 50P, 100K, 66Ca, 28Mg, 20S and micronutrients) as main plots and genotypes as sub-plots. Live and dead forage yield, shoot nutrient composition, and shoot nutrient uptake were measured at the end of the wet season (October 2000). Results and Discussion Initial application of high amounts of fertilizer did not improve forage yield of most of the genotypes compared with low fertilizer application (Table 59). This indicates that the initial application of high amounts of N and K fertilizer at the time of establishment had very little residual effects into the second year. At 15 months after establishment, live forage yield with low fertilizer application ranged from 0.27 to 2.39 t/ha and the high values of forage yield were observed with two germplasm accessions (CIAT 26110 and CIAT 26318) and one spittlebug resistant genetic recombinant, FM9503-S046-024 (Table 59). As expected, the performance of one of the parents, BRUZ/44-02 was very poor compared with other parents and genetic recombinants. One of the genetic recombinants, cv. Mulato (CIAT 36061), had more dead biomass with both levels of fertilizer application (Table 59). The gretest amount of total forage yield was obtained with one of the germplasm accessions, CIAT 26318 with both levels of fertilizer application. Two recombinants, cv. Mulato (CIAT 36061) and FM9503-S046-024 showed greater amount of dead biomass similar to one of the parents, CIAT 6294 with both levels of fertilizer application. These two recombinants were also superior in production of greater amount of green leaf and stem biomass (Table 60). Results on leaf and stem N content indicated that BRUZ/44-02 had greater amount of N per unit dry weight but its ability to acquire N (shoot N uptake) was lowest compared with other parents and genetic recombinants (Table 61). Shoot N uptake with low fertilizer application was greater for two accessions (CIAT 26110 and 26318), one parent (CIAT 6294) and one genetic recombinant (FM9503-S046-024). This genetic recombinant was also outstanding in its ability to acquire greater amounts of P, K, Ca and Mg from low fertilizer application when compared with parents, accessions and other genetic recombinants (Tables 62 and 63). Among the parents, B. brizantha CIAT 6294 was superior in P, K, Ca and Mg acquisition from low fertilizer application. It is important to note that live forage yield was associated with lower contents of not only essential nutrients but also Al in stems with both low and high fertilizer application. (Table 64). Live forage yield with low fertilizer application showed a significant negative relationship (-0.45**) with stem N content. This observation indicates that genotypes that are efficient in uitilization of N for the production of green forage is an important mechanism for superior performance with low fertilizer application. 113 Table 59. Genotypic variation as influenced by fertilizer application in live shoot biomass, dead shoot biomass and total forage yield of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 15 months after establishment (October 2000) LSD values are at the 0.05 probability level. Live shoot biomass Dead shoot biomass Total forage yield Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (kg/ha) Recombinants: BR97NO-0082 793 1125 385 427 1178 1552 BR97NO-0383 934 1376 375 516 1308 1892 BR97NO-0405 1230 1061 518 537 1748 1598 cv. Mulato (CIAT 36061) 1419 1824 1378 1650 2797 3474 CIAT 36062 1145 1355 415 761 1560 2116 FM9503-5046-024 2082 1712 1429 814 3511 2527 Parents: CIAT 606 907 1204 361 215 1267 1419 CIAT 6294 2022 2429 1030 1580 3052 4010 BRUZ/44-02 274 268 244 212 518 480 CIAT 26646 1194 1854 865 709 2060 2563 Accessions: CIAT 26110 2390 2231 1364 777 3755 3007 CIAT 26318 2379 2568 1618 1287 3996 3856 Mean 1397 1584 832 791 2229 2374 LSD (P=0.05) 800 812 1138 1195 1559 1745 Table 60. Genotypic variation as influenced by fertilizer application in leaf biomass, stem biomass and leaf to stem ratio of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 15 months after establishment (October 2000). LSD values are at the 0.05 probability level. Genotype Leaf biomass Stem biomass Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer (kg/ha) Recombinants: BR97NO-0082 578 741 215 384 BR97NO-0383 476 485 457 890 BR97NO-0405 629 523 601 538 cv. Mulato (CIAT 36061) 785 1014 634 810 CIAT 36062 808 1013 337 342 FM9503-5046-024 1356 1036 726 676 Parents: CIAT 606 571 779 335 425 CIAT 6294 1287 1184 735 1245 BRUZ/44-02 125 172 149 96 CIAT 26646 833 1185 361 668 Accessions: CIAT 26110 1371 1486 1019 744 CIAT 26318 1430 1412 948 1156 Mean 854 919 543 136 LSD (P=0.05) 440 428 415 522 114 Table 61. Genotypic variation as influenced by fertilizer application in leaf N content, stem N content and shoot N uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 15 months after establishment (October 2000). LSD values are at the 0.05 probability level. Leaf N content Stem N content Shoot N uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (%) (%) (kg/ha) Recombinants: BR97NO-0082 0.970 0.920 0.83 0.61 7.32 9.35 BR97NO-0383 1.050 0.960 0.91 0.62 8.75 9.61 BR97NO-0405 0.990 0.990 0.88 0.63 11.15 8.14 cv. Mulato (CIAT 36061) 0.740 0.830 0.86 0.56 10.88 13.06 CIAT 36062 0.880 0.910 0.65 0.53 9.08 11.06 FM9503-5046-024 0.980 0.900 0.90 0.54 19.86 12.53 Parents: CIAT 606 0.850 0.840 0.80 0.47 7.37 8.58 CIAT 6294 0.890 0.750 0.67 0.65 16.50 16.54 BRUZ/44-02 1.290 1.210 1.08 0.80 2.98 3.49 CIAT 26646 0.860 0.770 0.69 0.56 9.50 12.96 Accessions: CIAT 26110 0.900 0.720 0.40 0.51 16.41 13.75 CIAT 26318 0.770 0.700 0.65 0.48 16.99 14.98 Mean 0.930 0.870 0.76 0.58 11.40 11.33 LSD (P=0.05) 0.188 0.164 0.32 0.22 6.81 4.74 NS = not significant. Table 62. Genotypic variation as influenced by fertilizer application in leaf P content, stem P content and shoot P uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 15 months after establishment (October 2000). LSD values are at the 0.05 probability level. Leaf P content Stem P content Shoot P uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (%) (%) (kg/ha) Recombinants: BR97NO-0082 0.154 0.141 0.127 0.075 1.11 1.34 BR97NO-0383 0.149 0.160 0.088 0.072 1.08 1.30 BR97NO-0405 0.146 0.166 0.119 0.079 1.55 1.31 cv. Mulato (CIAT 36061) 0.153 0.149 0.118 0.068 2.04 1.98 CIAT 36062 0.169 0.190 0.110 0.074 1.73 2.20 FM9503-5046-024 0.179 0.155 0.118 0.076 3.12 2.07 Parents: CIAT 606 0.162 0.147 0.180 0.106 1.49 1.57 CIAT 6294 0.153 0.151 0.112 0.086 2.75 2.84 BRUZ/44-02 0.193 0.166 0.110 0.087 0.42 0.45 CIAT 26646 0.142 0.129 0.117 0.067 1.56 2.00 Accessions: CIAT 26110 0.126 0.132 0.060 0.098 2.33 2.42 CIAT 26318 0.126 0.135 0.101 0.075 2.69 2.64 Mean 0.154 0.151 0.113 0.080 1.82 1.87 LSD (P=0.05) 0.049 0.041 0.050 NS 1.06 0.78 NS = not significant. 115 Table 63. Genotypic variation as influenced by fertilizer application in shoot K uptake, shoot Ca uptake and shoot Mg uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 15 months after establishment (October 2000). LSD values are at the 0.05 probability level. Shoot K uptake Shoot Ca uptake Shoot Mg uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (kg/ha) Recombinants: BR97NO-0082 9.67 11.87 3.23 5.40 3.19 5.51 BR97NO-0383 11.97 9.49 2.79 4.64 3.06 4.71 BR97NO-0405 15.59 11.25 3.72 3.45 4.00 4.02 cv. Mulato (CIAT 36061) 15.27 21.76 4.96 6.08 5.93 6.41 CIAT 36062 16.59 13.73 4.69 6.31 6.13 8.45 FM9503-5046-024 27.52 18.32 9.48 7.07 10.27 9.07 Parents: CIAT 606 13.34 15.13 4.00 4.34 4.96 6.23 CIAT 6294 24.98 21.09 7.04 7.01 8.65 10.66 BRUZ/44-02 3.07 3.12 0.98 1.39 1.07 1.60 CIAT 26646 14.94 18.14 3.42 5.12 4.72 8.64 Accessions: CIAT 26110 23.11 19.92 6.40 7.06 7.35 11.24 CIAT 26318 28.61 18.65 6.65 7.10 9.25 12.06 Mean 17.05 15.46 4.78 5.50 5.71 7.51 LSD (P=0.05) 11.46 5.80 2.90 2.89 3.13 4.11 Table 64. Correlation coefficients ( r ) between green forage yield (t/ha) and other shoot traits of Brachiaria genotypes grown with low or high intial fertilizer application in a sandy loam oxisol in Matazul, Colombia. Shoot traits Low High fertilizer fertilizer Total (live + dead) shoot biomass (t/ha) 0.88*** 0.89*** Dead shoot biomass (t/ha) 0.56*** 0.59*** Leaf biomass (t/ha) 0.96*** 0.88*** Stem biomass (t/ha) 0.93*** 0.87*** Leaf N content (%) -0.31* -0.56*** Leaf P content (%) -0.31* -0.27 Leaf K content (%) -0.14 -0.31* Leaf Ca content (%) -0.16 -0.44** Leaf Mg content (%) -0.10 -0.02 Leaf Al content (%) -0.21 -0.37* Stem N content (%) -0.45** -0.43** Stem P content (%) -0.34* -0.33* Stem K content (%) -0.13 -0.17 Stem Ca content (%) -0.33* -0.46** Stem Mg content (%) -0.28 -0.30* Stem Al content (%) -0.30* -0.23 116 Conclusions Results from this field study indicated that the superior performance of the Brachiaria hybrid, FM9503- S046-024 at 15 months after establishment was associated with its ability to acquire greater amounts of nutrients from low fertility soil. 3.1.2.3 Field evaluation of promising hybrids of Brachiaria in the Llanos of Colombia Contributors: I. M. Rao, J. W. Miles, C. Plazas and J. Ricaurte (CIAT) Rationale Based on the data collected from greenhouse and field screening of a large number of Brachiaria hybrids, we selected 4 hybrids for further field testing to evaluate persistence with low nutrient supply in soil at Matazul farm of the altillanura. Materials and Methods A field trial was established at Matazul farm on 31 May this year. The trial included 4 Brachiaria hybrids (1251; 4015; 4132; 4624) along with 2 parents (B. decumbens CIAT 606 and B. brizantha CIAT 6294). The trial was planted as a randomized block in split-plot arrangement with two levels of initial fertilizer application (low: kg/ha of 20P, 20K, 33Ca, 14 Mg, 10S; and high: 80N, 50P, 100K, 66Ca, 28Mg, 20S and micronutrients) as main plots and genotypes as sub-plots with 3 replications. The plot size was 5 x 2 m. Results A number of plant attributes including forage yield, dry matter distribution and nutrient uptake are being monitored. 3.1.2.4 Screening accessions of Brachiaria humidicola for suppression of nitrification and nitrous oxide emission from soil Contributors: T. Ishikawa (JIRCAS, JAPAN), and I.M. Rao (CIAT) Rationale Ammonium-N is transformed into nitrite-N and nitrate-N by soil microorganisms (Figure 31), a process known as nitrification. Nitrification leads to substantial losses of applied fertilizer N through nitrous oxide (N2O) emission and runoff/leaching losses of nitrate from agricultural production systems. This is often associated with nitrate pollution of ground water and aquatic bodies. Nearly 50 to 70% of the applied fertilizer N is lost because of nitrification, causing enormous environmental pollution problems and also inefficiency in N utilization. Nitrous oxide, one of the greenhouse gasses, is emitted from the soil because of nitrification. Preliminary estimations indicate that N2O emissions from the fertilizer N range from 7.3% of applied N for field crops such as maize to 12.0% of applied N for grasslands. By controlling nitrification in soils, it will be possible in future to reduce N fertilizer inputs into agricultural production systems and also minimize nitrate pollution in aquatic systems and ground water. 117 We found that a tropical grass, Brachiaria humidicola that is widely adapted to lowland agroecosystems (savannas) of humid and subhumid tropics, particularly in South America has the ability to suppress nitrification in soil and emission of N2O to the atmosphere. Suppression of nitrification by Brachiaria humidcola Brachiaria humidcola N2O Greenhouse effect ~ Absorption by plants Ammon~ium-oxidizing Nitrite-oxidizing Bacterium bacterium NH + 4 ~ NO - 2 NO - 3 Suppression of nitrification Nitrification Leaching Figure 31. Mechanism of nitrification and nitrification suppression by Brachiaria humidicola Materials and Methods Three tropical grasses, Brachiaria decumbens, Brachiaria humidicola and Melinis minutiflora were grown in Wagner pots filled with Typic Hapludands [pH (H2O) 6.0 and C/N ratio 11.68)]. Plants were grown in growth chamber with day/night temperature regimes of 30 ºC and 28 ºC, respectively with a 14 h photoperiod. Six weeks after sowing (WAS), 1.422 g of ammonium-nitrogen was supplied as ammonium sulfate. At 8 weeks after sowing, plants were separated from the soil. Root exudates were collected from these plants by keeping the plants in distilled water (1L pot-1) for 24 h, plants were then harvested and dried for dry weight measurements. Soils were sampled for chemical analysis and for the initiation of nitrification study that followed. The effect of root exudates on the multiplication of ammonium oxidizing bacteria (AOB) was measured by MPN (most probable number) method. Nitrification Study: Nitrification experiment was initiated to test the residual effect of the three grasses on soil nitrification during a 24 day period. Ammonium-N was applied (1.422 g NH4 SO4) to each of the sampled soil (2.3 kg pot-1) and incubated for 24 days. Soil water content was maintained at 50% water holding capacity during this period (i.e. 24 day period after the harvest of plants at 8 WAS). Soil extract was made and its effect on multiplication of AOB was measured by MPN method. Soil samples were collected at various intervals and analyzed for NO3 and NH4 forms of nitrogen. Nitrous oxide emission from sampled soil was monitored by collecting the air samples at periodic intervals and N2O levels were analyzed by gas chromatography. 118 Results and Discussion Ammonia oxidizing bacteria (AOB) were nearly 10 times higher in soils where B. decumbens and Melinis minutiflora were grown at 8 WAS compared to soils where B. humidocola was grown (10,000 vs 1000 g-1 dry soil). Residual effect of B. humidicola on suppression of AOB lasted for about 12 days after the plants were harvested, but subsequently the AOB began to increase and reached levels similar to B. decumbans and M. minutiflora treatments at 24 days after plants were harvested (results not shown). For nitrite- oxidizing bacteria, however there were no significant differences among B. decumbans, B. humidicola and M. minutiflora treatments (results not shown). Results on ammonium-N in soils were presented as the percentage of initial amount of ammonium-N applied (Figure 32). Nearly 50% of the ammonium-N was lost by 12 days after the initiation of nitrification treatment in soils where B. decumbens and M. minutiflora were grown (Figure 2). However, in soils where B. humidicola was grown, there was no significant change in ammonium-N levels up to 12 days but subsequently declined. By 24 days after the nitrification study was initiated, NH4-N in soils of B. humidicola treatment was similar to that of B. decumbens and M. minutiflora. Most of the applied ammonium-N was converted into nitrate-N or was lost as N2O in all the three treatments by 24 days after the plants were harvested. Thus, the residual effect of B. humidicola on AOB has lasted only for about 12 days after the plants were harvested. 120 100 Bd Mm Bh 80 60 40 20 0 0 4 8 12 16 20 24 Days after ammonium application Figure 32. Percent of intial amount of ammonium-N in soil in relation to days after ammonium application Nitrous oxide emission during the nitrification study was substantially higher for B. decumbans and M. minutiflora (31.0 and 29.3 µg-N m-2 hr-1) compared to B. humidicola (5.0 µg-N m-2 hr-1) treatment (Figure 33). For B. decumbans, and M. minutiflora treatments, N2O emission reached the highest levels between 8 and 12 days after the initiation of nitrification study, which is similar to control pots (i.e. no plants). Soil 119 NH4-N (% of initial amount) extracts and root exudates of B. humidicola treatment suppressed AOB, whereas no such effect was observed for B. decumbans or M. minutiflora treatments (results not shown). 8 Control 7 No Plants Bd 6 Bh Mm 5 4 3 2 1 0 0 4 8 12 16 20 -1 Days after ammonium application Figure 33. Nitrous oxide emission from soil in relation to days after application of ammonium. Control pots received no ammonium application while pots with no plants received ammonium application. Conclusions Our results strongly support the notion that B. humidicola has the ability to suppress nitrification by inhibiting the biological activity of ammonium oxidizing bacteria (AOB) in the soil. This was demonstrated by substantial decrease in AOB populations in soils where B. humidicola was grown. Also, nitrous oxide emissions, which is an indicator of the AOB biological activity was very low for B. humidicola treatment. Our results support the hypothesis that B. humidicola suppress nitrification and N2O emission by inhibiting the activity of AOB in the soil. This may be achieved by secreting organic compounds from the roots that have the inhibitory effect on these AOB bacteria. Also, we have demonstrated that the residual effect of this tropical grass on nitrification of the soil will be about 24 days. The Genetic Resources Unit of CIAT has a germplasm collection of about 62 accessions of Brachiaria humidicola. We plan to evaluate these germplasm accessions of B. humidicola in order to identify genotypic differences in their ability to suppress nitrification. This work is expected to contribute toward identification of germplasm accessions with greater ability to inhibit nitrification process in soil. The selected accessions could then be used in breeding programs for developing genetic stocks that combine greater forage production potential with high levels of nitrification inhibition capacity. Currently we are also working on identification of the organic compound/s that are responsible for this unique property of nitrification inhibition in the root exudates of Brachiaria humidicola. Also, efforts are underway to understand the mechanisms of nitrification inhibition in these root exudates of B. humidicola. 120 N2Oemission (µg-N m-2 h-1) 3.1.3 Differences in phosphorus acquisition from less available phosphorus forms in an oxisol as determined by isotope exchange kinetics Contributors: S. Buehler, A. Oberson, E. Frossard (ETH, Zurich, Switzerland), D. Friesen (IFDC- CIMMYT, Nairobi, Kenya) and I. M. Rao (CIAT) Rationale The ability to grow on low P soils differs widely between plant species and even among varieties. These differences have been attributed to several strategies for optimizing P uptake or P use efficiency. The strategies for P uptake improvement include root system morphology, root hair density, symbiosis with mycorrhizal fungi and modification of the rhizosphere by root exudates or phosphatases to access P forms of low availability. Previous research showed that Arachis pintoi accesses sparingly soluble P sources by high response to fertilization with Al-phosphate, Ca-phosphate or organic phosphate (phytic acid). Brachiaria species are reported to be well adapted to low P acid soils with the variety Brachiaria decumbens CIAT 606 being planted on over 40 million ha of low P acid soils in Latin America. Adaptation of Brachiaria species to low-P soils is mainly attributed to the extensive fine root system, mycorrhizal association, and lower internal requirement of P for plant growth together with enhanced secretion of phytase under low P conditions (e.g., Brachiaria decumbens). Infertility of acid tropical soils is caused by multiple stress factors, among others Al-toxicity. Brachiaria decumbens also tolerates high Al-concentrations in the soil solution, possibly due to intracellular complexation of Al-ions with organic acids. Plants with special P uptake mechanisms may contribute to more efficient soil P use or could increase the recovery of applied fertilizer if they take up P that is normally not available to other plants. This might reduce the need of high P fertilizer inputs, although on the long-term the use of germplasm with special P uptake mechanisms would lead to soil P mining. A strategy to contribute to agricultural sustainability would be to minimize the P fertilizer requirement needed to produce an economic return through the use of more efficient crop and forage germplasm. Differences between plants in the accessed P forms can be detected by growing plants on soils labeled with radioactive P isotopes and comparison of their isotopic composition. The L value is derived from the isotopic composition of the P taken up by the plant and expresses the total amount of soil P, which is potentially available to the plant. This amount can be compared to the E value of the same soil, which is based on the isotopic composition, measured or extrapolated, in the solution of a labeled soil suspension after a defined time of isotopic exchange. It was shown for a large range of soils, L values determined with common bentgrass (Agrostis capillaris) and E values calculated for the same time of isotopic exchange are not significantly different, and therefore Agrostis capillaries takes up only isotopically exchangeable P. In contrast, significant differences between E values and L values on the same soil for the same time of isotopic exchange are interpreted by plant P uptake from non-exchangeable P pools. Different L values of different plants on the same soil indicate that these plants do access P pools with different isotopic exchangeability. Higher L values were attributed to special P uptake mechanisms, especially exudation of organic acids, as citric acid in the case of lupine or piscidic, malonic or oxalic acid in the case of pigeon pea. Seed P is an important factor affecting the isotopic composition of the P taken up by the plant and with this L values, especially under P-limitation and with little plant P uptake. Adding carrier-P, i.e. a simultaneous 31P addition with the radioactive label, could increase plant growth and P uptake. It is assumed that L values are independent of the amount of carrier added to the soil. Besides the effect of enhancing plant growth, the application of carrier with the 33P label is recommended to avoid the fixation of the label. 121 As part of a Ph.D. thesis study funded by a special project from SDC, Switzerland, two key forage species (Arachis pintoi and Brachiaria decumbens) were compared with three low-P adapted crop cultivars (maize, bean and upland rice) in terms of their ability for P uptake from sparingly available P pools based on their L values. A soil was chosen with very low available P (determined with Bray II) in order to guarantee P limitation for plant growth. The L value of Agrostis capillaris was determined as reference assuming that this plant does not access any non isotopically exchangeable P. E values were determined in a batch experiment without carrier application to determine the isotopically exchangeable soil P. L values were determined in two experiments, one without and one with a P carrier application of 10 mg P kg-1 soil. This amount was chosen as, with the application of the same amount to a similar soil, Bachiaria species and Arachis pintoi increased biomass production and P uptake but remained P limited. Materials and Methods The soil chosen (well-drained oxisol) for this study was cultivated as improved grass legume pasture starting with rice in 1993, with under sown pasture. The procedure to determine E values is based on the measurement of the specific activity (33PO4/ 31PO4) of phosphate ions in the soil solution after an addition of carrier free 33PO4 in a soil-solution system at steady state. The isotopically exchangeable P (Et) was calculated assuming that, at any given exchange time, the specific activity of phosphate in solution is equal to the specific activity of the exchangeable phosphate on the solid phase: r t R [Eq. 1] = 10*Cp E t or: 10C [Eq. 2] E t = R × p r t and rt/R is extrapolated as: −n  ( 1 )  [Eq. 3] r t n = r1 t + r1  + r∞ R R  R  R   where R is the introduced radioactivity in MBq ml-1and rt is the radioactivity remaining in the solution after t minutes. The other parameters can be determined experimentally: n is a parameter calculated as the slope of the linear regression between ln(rt/R) and ln(t) for t≤100 minutes of the measured values, r1/R is the interception of the regression when t=1, and r∞/R is calculated as: r ∞ = 10 * C p R P i Where Pi is total inorganic P and the ratio r∞/R repr at infinite time. L value determination. The experimental condition L values are summarized in Table 2. The cultivars u 122[Eq. 4]esents the radioactivity remaining in the soil solution s of the two pot experiments carried out to determine sed were Brachiaria decumbens (CIAT 606), Arachis pintoi (CIAT 18744) and rice (Oryza sativa var Savanna-6) in experiment one. Additionally we used beans (Phaseolus vulgaris AFR 475) and maize (Zea mays NST 90201(s) co-422-2-3-1-7-2-1), an inbred line derived from a triple hybrid developed by the Thai Department of Agriculture, selected as tolerant to low-P conditions. In both cases common bentgrass (Agrostis capillaris) was grown as control plant without adaptation to low P conditions. The soil was labeled by adding the quantities of 32PO or 33 4 PO4 ions in 10 mL water to portions of 1.5 kg incubated soil and were thoroughly mixed to ascertain an even distribution of the isotope (details given in Table 65). Table 65. Soil (0-15 cm) properties determined on air-dried samples (except bulk density). Total P (mg kg-1) 242 Total Pi (mg kg-1) 86.4 Resin P (mg kg-1) 1.5 Bray-II P (mg kg-1) 3.1 pH (in H2O) 4.3 Total C (g kg-1) 23.7 Total N (g kg-1) 1.6 Aluminum-saturation ( %) 68 Bulk density (g cm-3) 1.3 Agrostis was grown from 100 mg of seeds (corresponding to about 800 seeds) in both experiments, which were sown directly into each pot. All other plants were pregerminated on filter paper before planting into the pot at numbers indicated in Table 66. The pots with beans and Arachis pintoi were inoculated with a suspension of the Rhizobium strains CIAT 899 and CIAT 3101, respectively. During the experiment soil moisture was controlled by weighing and kept at 50 % of the water holding capacity. Table 66. Experimental conditions in pot experiment 1 and 2 for determination of L values. Experiment 1 Experiment 2 Plant species, quantities of soil Arachis pintoi, 2 kg soil, 2 plants Arachis pintoi, 0.9 kg soil, 1 plant and plants per pot Brachiaria decumbens, 2 kg soil, 2 Brachiaria decumbens, 0.9 kg soil, 2 plants plants Rice, 2 kg soil, 2 plants Rice, 0.9 kg soil, 2 plants Beans, 0.9 kg soil, 1 plant Maize, 3.4 kg soil, 1 plant Agrostis capillaris, 500 g, 100 mg Agrostis capillaris, 400 g soil, 100 mg seeds seeds Labeling 32PO4, 5.2 MBq kg-1 soil 33PO4, 3.7 MBq kg-1 soil Carrier none 10.26 mg P as KH2PO -1 4 kg soil, applied with labeling solution Replicates 4 5 Location greenhouse, CIAT, Colombia Biotron, ETH, Switzerland Experimental conditions Maximum light intensity 16 h daylight, light intensity ~ 300 µ ~ 1100 µ mol m-2 s-1 mol m-2 s-1 Temperature 38/20 °C (max/min d/n, over whole 24/20 °C (constant) growth period) Humidity 90/40 % (max/min) 65 % (constant) Duration of plant growth: 2 months 11 weeks 123 After two months or eleven weeks, respectively, plant shoots were harvested and dry matter was weighed after 48 h drying at 80° C. About 200 mg of a homogenized sample of the whole shoot biomass in the first or half of the total shoot in the second experiment, cut in pieces <2mm, was calcinated at 550° for 4 hours. Plant P content (p) was determined after solubilization of the ash in 1-5 mL of 11.3 M HCl. The same method was used for the determination of the seed P content, measuring ball-milled samples of 100 mg (Agrostis capillaris), two seeds (Arachis pintoi, rice, beans and maize) or five seeds (Brachiaria decumbens), with five replicates each. The plant 33P (r) content was measured by scintillation counting of diluted (to avoid quench effect) samples using a liquid scintillation analyzer (Packard 2500 TR) and Packard Ultima Gold scintillation liquid. The measured radioactivity was decay corrected back to the day of soil labeling. The L values, expressed as mg P kg-1 soil, were calculated with the P-concentrations and activities measured in the total shoot. Experiment 1: Without carrier: L = R * p [Eq.5] r Experiment 2 With carrier:  L = Q R * p  Q r −1 [Eq.6]  *    The source of P taken up by the plant in the experiment with carrier addition can be calculated as: Psoil = p - P [Eq.7] carrier P = Q * r carrier R [Eq.8] where R is the quantity of 33PO or 32 4 PO4 used to label exchangeable soil P (MBq kg-1 soil) and Q the quantity of carrier added (mg P kg-1 soil), r is the quantity of 33PO or 32 4 PO4 (MBq kg-1 soil) and p is the quantity of 31PO4 (mg kg-1 soil) in the plant shoots. Pcarrier and Psoil are the total amount of P derived from the carrier solution or from soil respectively. However, the P content of the seed is a third P source and uptake from this source could not be distinguished from the P taken up from soil. Therefore, Psoil is actually the sum of the P taken up from soil and from the seed, and the specific activity of the P taken up from soil is diluted. This results in an overestimation of the L value in both experiments. To increase the accuracy of the L value, the following correction was applied: L L p th = ( ) [Eq.9] p + Pseeds 124 where Lth is the corrected value, L the value calculated with Eq. 5 or 6 and Pseeds the P content of the sown seeds per pot. Another possibility to correct for the seed P influence is to subtract the total seed P content from plant P uptake for the calculation of the L value: This correction assumes that 100 % of seed P was taken up by the plant and allocated to the shoot. Therefore, it corrects for the highest possible influence of seed P. R ( p - P ) [Eq. 10] L = seed r Acid phosphatase activity determination: Three bulk soil samples were taken at random after harvesting plants in all pots, air dried and roots were removed carefully by sieving soil at 2 mm. Acid phosphatase activity at pH 6.5 of soil samples derived from the planted pots and soil incubated without plants at the same conditions was measured using 1 g air-dried soil. Statistical Analysis: The effect of plants in the pot experiment and the effect of the experimental conditions on parameters of isotopic exchange in the E value determination were tested by analysis of variance (ANOVA). If the F-test was significant (P<0.05), the means were compared using Tukey's multiple range test. Results and Discussion L values determined without carrier and correction for seed P influence: The biomass production of all plants in the first experiment was very low and the total P uptake was hardly higher than the P content of the seeds (Tables 68 and 69). Correction for the contribution of seed-P according to Eq. 9 resulted in a marked reduction of the uncorrected L values (Table 68). It is, however, doubtful whether this correction, which was established for L value determination with Agrostis capillaris and Lolium perenne as model plants on sand culture, is also valid for other test plants and for all soil types. The correction with Eq. 10 was only applicable in the case of Brachiaria decumbens as for all other test plants Pseed>p. Consequently, the corrected L values may rather give an impression of the order of magnitude of the seed P influence than represent exact values. Most studies comparing L values of different plants on low P soils, may have underestimated the problem of seed P influence. As the P reserves in the seed are in most cases relatively high in comparison to the P taken up from soil, the L value can not be calculated without correction for seed P uptake. Due to the uncertain influence of seed-P, the interpretation of the L values remains limited. Additionally, the L value of Agrostis capillaris could not be used as reference as the ratio of plant P uptake to seed P was very low, too. However, in the case of Brachiaria decumbens with the smallest influence from seed P (Table 68 and 69), the corrected Lth-value remains much higher (131 mg kg-1 with Eq.9 or 127 mg kg-1 with Eq.10, respectively) than the extrapolated E8weeks-value of 64 mg kg-1 determined for the same soil (Table 67). As Eq. 10, with the subtraction of total seed P from P export in the plant shoot, corrects for the highest theoretically possible influence of seed P, the L value of Brachiaria decumbens indicates that P additional to the isotopically exchangeable P was taken up. However, it should be mentioned that the extrapolation of E values on such very low P soils is difficult and the precision of the calculated E8weeks is therefore limited. On the other hand, the L value of Brachiaria decumbens is also higher than the total soil Pi extracted with the sequential P fractionation (Table 65). This fact reinforces the assumption that organic P or very recalcitrant P forms contributed to the P uptake of Brachiaria decumbens. 125 Table 67. Parameters of isotopic exchange of the used soil. r1/R†1 0.03 cp ‡ 0.003 mg l-1 n¶ 0.43 E1 # 1.1 mg kg-1 E -1 8weeks# 64 mg kg † ratio of radioactivity remaining in soil solution to radioactivity added at time 0 after 1minute of isotopic exchange ‡ P concentration in the soil solution measured at soil:water ratio 1:10 ¶ Parameter of isotopic exchange describing the decrease of radioactivity in the soil solution #Quantity of P exchangeable within 1 minute or within 8 weeks (calculated with Eq. 3) Table 68. Biomass production, P uptake and L values of the compared plants in experiments 1 and 2. Plant Shoot dry weight P uptake Shoot P L† Lth‡ Material concentration (g per pot) (mg per pot) µg g-1 dry weight (mg P kg-1 soil) Exp 1 Exp 2 Exp 1 Exp 2 Exp 1 Exp 2 Exp 1 Exp 2 Exp 1 Exp 2 A. pintoi 1.6a 2.4b 0.9a 2.1b 561ab 877b 185a 4.0 46b 3.1 B. decumbens 0.3bc 1.9bc 0.22b 1.1b 729a 581c 153ab 0.9 131a/ 0.9 127§ Rice 0.6b 2.3b 0.25b 1.1b 417b 478cd 125b 1.4 39b 1.1 Maize - 6.3a - 3.9a - 622c - 4.7 - 3.8 Beans - 1.0c - 1.5b - 1350a - 1.7 - 1.1 Agrostis 0.2c 1.0c 0.14b 0.4b 697ab 392d 128ab 3.3 6.7b 1.6 capillaris ANOVA *** *** *** *** * *** * n.s.¶ *** n.s. *,*** Significant at the 0.05 or 0.001 probability level, respectively. Values within columns followed by the same letter do not differ significantly (P=0.05) according to Tukey's test. †L value without seed-P correction ‡L value with the seed -P correction, Eq. 9 § second value: corrected with seed P correction, Eq. 10 ¶ not significant Table 69. Average seed weight and seed P content of the used varieties Plant material Weight per seed Total P in sown seeds (mg) (µg) Arachis pintoi 158 1200/ 2 seeds Brachiaria decumbens 4.6 26.7/ 2 seeds Rice 44 282/ 2 seeds Maize 306 900/ 1 seed Beans 174 540/ 1 seed Agrostis capillaris 0.125 480/100 mg seeds The adaptation of Brachiaria species to low P soils is mainly attributed to soil exploration by an abundant fine root system and mycorrhizal association. In addition, it was shown in a pot experiment with different 126 added P sources, that Brachiaria dictyoneura cv. Llanero can acquire P from less available inorganic (aluminum phosphate, as AlPO4) and organic (phytic acid) forms. Acid phosphatase activity in roots of Brachiaria dictyoneura was increased with decreasing soil P supply, and Brachiaria decumbens grown under low P condition in nutrient solution was shown to secrete the highest amount of phytase in comparison to 15 other plant species. In our study acid phosphatase activity measured in the pot soil samples, and in turn the potential to mineralize available phosphomonoesters, was only significantly increased (p<0.001) for Arachis pintoi in the first experiment and was increased significantly (p<0.001) for all plants but Agrostis capillaris in comparison to the control soil without plant in the second experiment (Table 70). However, as the measurements were not restricted to rhizosphere soil, local effects in that zone would not have been detected. Table 70. Phosphatase activity in soil samples derived from pots after plant harvest. Plant species Phosphatase activity Exp. 1 Exp. 2 __________µg nitrophenol g-1 h-1________ Arachis pintoi 426a 322a Brachiaria decumbens 285b 342a Rice 295b 295ab Maize - 332a Beans - 286b Agrostis capillaris 236b 242c Control 219b 225c ANOVA *** *** ***Significant at the 0.001 probability level, values within a column followed by the same letter do not differ significantly (P=0.05) according to Tukey’s test. The influence of carrier application: As the correction for seed P influence was difficult, the L value determination without carrier application was unsatisfying for the tested plants, with exception of Brachiaria decumbens. To overcome the difficulties of small total P uptake and biomass production, the second experiment was carried out with the application of KH2PO4 (10.3 mg P kg-1 soil) as a carrier with the labeling solution and the duration of plant growth was extended from two month to eleven weeks and smaller pots were used to reach higher biomass production and a higher soil exploration by the roots. The application of a P carrier resulted in much smaller L values (mean of all plants 2.7 mg P kg-1 soil) than without carrier (mean 148 mg kg-1) and there were no significant differences between plants. One possible explanation of the difference found between L values determined with or without carrier application is that an application of 10 mg P kg-1 to a soil with a very low P concentration in the solution (in this case approximately 3 µg l-1) could have a high impact on the processes in this system. Instead of isotopic exchange, a net diffusion process may dominate and sizes of pools are changed. High influences of carrier application on E values, especially for high P sorbing soils, were found before and were explained by the influence of carrier P on the process of isotopic exchange as well as by the fact of 32PO4 fixation. Additionally to the carrier application also the different experimental conditions, especially the smaller pot size, might have influenced the L-value. However, a smaller soil volume and therefore higher root exploration and biomass production per kg soil should, if at all, lead to an increase of L values by higher 127 root activity and P mobilization and not to a decrease. In our study, the nearly identical values of the specific activities of the plants and the applied carrier indicate that the carrier P was the main source for the P taken up by the plant. Separation of the P sources using Eq. 7 and 8 shows that on average 81% of P taken up by the plant derived from the carrier (Table 71). Of the 19 % of the total plant P uptake derived from another source a part is actually seed P. Therefore it can be assumed that almost no soil P was taken up and that the application of carrier is not valid for the determination of L values on low-P highly P sorbing soils. Table 71. Amount of P derived from applied carrier and percentage of P derived from other sources in Experiment 2 (calculated with Eq. 7 + 8). Plant species Total P uptake in P derived from P derived from other sources plant shoot carrier (soil and seed) (mg per pot) (%) Arachis pintoi 2.1b 1.5b 27a Brachiaria decumbens 1.1b 1.0b 9b Rice 1.1b 0.9b 16b Maize 3.9a 2.6a 26a Beans 1.5b 1.0b 13b Agrostis capillaris 0.4b 0.3C 24a (average) 19 ANOVA *** *** *** *** Significant at the 0.001 probability level, values within columns followed by the same letter do not differ significantly (P=0.05) according to Tukey's test. Conclusions A higher L value than E value for Brachiaria decumbens suggested P uptake from less available P forms from a low P soil. For all other plants, the contribution of the seed P to plant P uptake did not allow the calculation of exact L values. Therefore, drawing conclusions about the access of different P pools by different plants was not possible. L values determined with or without carrier P differed widely and suggested that a carrier application is not recommendable for using soils with very low P supply. Results from this study indicate that it is possible to use L value determination as a screening method to identify the most promising Brachiaria hybrids with adaptation to low P supply in soil. Further research work is needed to identify specific physiological and biochemical mechanisms contributing to the ability of B. decumbens to acquire P from less aavailable forms from low P oxisol. 3.1.4 Studies on genotypic variation in Arachis pintoi for tolerance to low phosphorus supply Contributors: N. Castañeda, N. Claassen (University of Goettingen, Germany); A. Betancourt, R. Garcia, J. Ricaurte and I. M. Rao (CIAT) Rationale For the last two years, we reported the progress from the field and greenhouse studies that were aimed to determine genotypic differences among ten accessions of Arachis pintoi in P-acquisition and utilization from low P soil. This year we report the progress made on persistence of these ten accessions. 128 Materials and Methods A field study is in progress at “La Rueda” ranch, Montañita, Caquetá (latitude 1º 25’ N, longitude 75º 27’ W and 180 m.a.s.l). Plant growth was monitored since June 1998. The mean rainfall, temperature and relative humidity were 3500 mm/year, 25ºC and 75% respectively. The experiment was laid down in a split plot RCBD with three P levels [native P (NP), phosphate rock (PR), triple super phosphate (TSP)] as main plots and ten genotypes [CIAT 17434 (commercial), 18744, 18748, 22159, 18745, 18751, 22160, 18747, 22155, 22172] as subplots. The experiment was replicated three times. Application (kg P ha-1) of PR and TSP was at 50 and 20, respectively. Plants were harvested at 32 months after establishment. Results and Discussion Last year, we reported that CIAT 22159 was outstanding in terms of persistence during second year as determined by leaf area index, shoot biomass and shoot P and Ca uptake. But this accession was relatively slow in establishment. At 32 months after establishment and after a short dry spell, we evaluated the persistence of the same 10 accessions with 3 different sources of P applied at the time of establishment. Initial application of PR and TSP at establishment had no residual effect on the performance of 10 accessions. (Table 72). Live forage yield was greater for CIAT 22159 and CIAT 18744 with no P application treatment. This was not due to better leaf production but due to a large biomass of stolons (Table 73). With no P application treatment, CIAT 22159 which was reported to be better for persistance in association with aggressive grass, B. dictyoneura cv. Llanero at Carimagua was also superior in its ability to produce greater leaf area development and therefore leaf biomass production. Dead forage yields were greater with this accession with no P application. Table 72. Influence of P fertilizer source on genotypic differences in leaf area index, live forage biomass and dead forage biomass of Arachis pintoi grown a low P soil at Montañita, Caquetá. Measurements were made at 20 months after planting. Plant P source Accession of A. pintoi LSD0.05 Attributes 17434 18744 18745 18747 18748 18751 22155 22159 22160 22172 Leaf area NP 1.34 0.78 0.58 0.68 0.99 0.65 0.76 1.29 0.71 0.71 0.57 index RP 0.89 0.75 0.72 0.44 1.10 0.81 0.69 0.78 0.63 0.77 0.48 (m2/m2) TSP 1.19 0.78 0.76 0.56 1.06 0.93 0.58 0.71 0.87 0.93 0.49 Live forage NP 2.22 3.08 1.39 2.32 1.93 2.15 3.07 3.54 1.74 2.84 1.14 yield RP 2.31 2.38 1.60 1.86 2.83 2.33 2.19 3.39 2.15 2.40 0.91 (t/ha) TSP 2.04 3.14 1.41 2.18 2.35 2.12 1.41 2.44 1.79 2.84 1.14 Dead forage NP 1.03 1.55 0.96 1.05 1.11 1.52 1.71 1.80 1.22 0.99 0.65 yield RP 1.39 1.92 1.16 1.23 1.40 1.34 0.84 1.22 1.12 0.76 0.64 (t/ha) TSP 1.47 1.23 1.60 1.31 1.18 1.19 0.55 1.16 2.24 1.12 0.77 NP= Native phosphorus (0 kg P/ha); RP= Rock phosphate (50 kg P/ha); TSP= Triple super phosphate (20 kg P/ha). 129 The differences among the 10 genotypes in terms of leaf N and P contents were small when compared with leaf biomass production (Table 73). The content of N and P in green leaves was greater with the accession CIAT 18751. Nutrient uptake, particularly P and N by green leaves during the dry spell was not significantly different among 10 acessions for no P treatment. Among the ten accessions, CIAT 17434, the commercial cultivar was outstanding in its ability to acquire Ca particularly from no P treatment (Table 74). Table 73. Influence of P fertilizer source on genotypic differences in leaf biomass, leaf N content and leaf P content of Arachis pintoi grown a low P soil at Montañita, Caquetá. Measurements were made at 20 months after planting. Plant P source Accession of A. pintoi LSD0.05 Attributes 17434 18744 18745 18747 18748 18751 22155 22159 22160 22172 Leaf NP 629 501 327 417 578 406 498 541 423 423 235 biomass RP 518 431 388 258 666 576 417 477 389 438 267 (kg/ha) TSP 562 505 409 329 568 548 361 431 522 564 NS Leaf N NP 24.9 32.6 27.8 32.8 26.8 32.6 30.1 29.5 30.5 28.5 NS content RP 30.2 36.2 36.4 38.1 35.1 34.9 29.5 31.5 33.9 30.9 4.4 (g/kg) TSP 32.5 34.8 40.0 35.0 32.0 36.3 30.0 33.2 33.6 28.3 6.9 Leaf P NP 2.1 2.2 2.5 2.4 2.1 2.4 2.4 2.1 2.1 2.0 NS content RP 2.1 2.7 2.8 2.8 2.4 2.6 1.9 2.3 2.3 2.2 0.4 (g/kg) TSP 2.4 2.5 2.7 2.8 2.1 2.7 2.0 2.3 2.7 2.0 NS NP= Native phosphorus (0 kg P/ha); RP= Rock phosphate (50 kg P/ha); TSP= Triple super phosphate (20 kg P/ha). Table 74. Influence of P fertilizer source on genotypic differences in P, N and Ca uptake by green leaves of Arachis pintoi grown a low P soil at Montañita, Caquetá. Measurements were made at 20 months after planting. Plant P source Accession of A. pintoi LSD0.05 Attribute 17434 18744 18745 18747 18748 18751 22155 22159 22160 22172 NP 1.30 1.10 0.84 0.98 1.22 1.02 1.05 1.16 0.90 0.85 NS P uptake RP 1.11 1.15 1.06 0.69 1.57 1.47 0.82 1.09 0.86 0.95 0.59 (kg/ha) TSP 1.37 1.24 1.08 0.92 1.19 1.47 0.66 1.00 1.37 1.13 0.60 NP 15.6 16.58 9.23 13.99 15.32 13.65 15.66 15.99 12.92 12.12 NS N uptake RP 15.77 15.50 13.98 9.72 23.37 20.10 12.52 15.11 12.50 13.47 9.2 (kg/ha) TSP 18.77 17.54 15.98 11.77 18.43 19.85 10.63 14.39 17.09 15.97 NS NP 8.25 4.17 2.55 3.50 3.95 3.73 3.67 5.33 3.74 5.00 3.73 Ca uptake RP 7.07 4.05 4.07 2.92 5.34 4.66 3.70 6.51 2.91 5.31 3.32 (kg/ha) TSP 5.72 3.23 4.34 3.17 4.32 3.71 3.53 3.87 3.71 5.83 NS NP= Native phosphorus (0 kg P/ha); RP= Rock phosphate (50 kg P/ha); TSP= Triple super phosphate (20 kg P/ha). Conclusions It appears from this field study that CIAT 22159 may be a better accession in terms of persistence with no P fertilizer input. 130 3.1.4.1 Field evaluation of most promising accessions of Arachis pintoi in the Llanos of Colombia Contributors: I. M. Rao, M. Peters, C. Plazas and J. Ricaurte (CIAT) Rationale Last year, we reported the progress from the field studies carried out at Caqueta that were aimed to determine genotypic differences among ten accessions of Arachis pintoi in persistence with low P supply in soil. Based on these data and the data collected from multilocational evaluation, we have assembled a set of 8 genotypes for further testing at two sites (Piedmont and Altillanura) in the Llanos of Colombia. The site in Piedmont is close to La Libertad (CORPOICA Experimental Station) and the soils in this region are relatively more fertile than in the Altillanura. The site in Altillanura is at Matazul farm where the soils are relatively infertile (sandy loam). Materials and Methods Two field studies were established during May this year. The trial in Piedmont was planted as monoculture while the trial in Altillanura was planted in association with a grass. This is based on the expected end use of the legume. We expect multiple use for this legume in the Piedmont area (e.g., cover legume in plantations). The trial in the Piedmont included 8 accessions of Arachis pintoi (CIAT 17434; 18744; 18747; 18748; 18751; 22159; 22160 and 22172). The trial in the Altillanura included 4 accessions (CIAT 17434; CIAT 18744; CIAT 18748 and CIAT 22159) planted in association with Brachiaria decumbens. Both trials were planted as randomized block in split-plot arrangement with two levels of initial fertilizer application (low: kg/ha of 20P, 20K, 33Ca, 14 Mg, 10S; and high: 80N, 50P, 100K, 66Ca, 28Mg, 20S and micronutrients) as main plots and genotypes as sub-plots with 3 replications. Results At 140 days after establishment of the trial in Piedmont, CIAT 18751 was found to be outstanding in its ability to establish rapidly. Two accessions (CIAT 17434 and 18751) were responsive to high level of fertilization while one of those two accessions (CIAT 18751) together with CIAT 18747 were among better performers with low fertilizer application as determined by soil cover. A number of plant attributes including forage yield, dry matter distribution and nutrient uptake are being monitored for both trials. Activity 3.2 Genotypes of grasses and legumes with dry season tolerance Highlights • Showed that the superior performance of the Brachiaria hybrid, FM9503-S046-024 which maintained greater proportion of green leaves during moderate dry season in the llanos of Colombia, was associated with lower levels of K and N content in green leaves. • Field screening of 16 genotypes of Brachiaria and 13 accessions of Arachis pintoi with long dry season in Costa Rica resulted in identification of Brachiaria hybrid CIAT 36061 as not only adapted to drought but also superior in maintaining greater level of nitrogen (crude protein) in green leaves. 131 Progress towards achieving milestones • Brachiaria accessions and hybrids with superior tolerance to drought relative to commercial cultivars identified. One hybrid of Brachiaria (FM9503-S046-024) was identified as promising material for areas with moderate drought stress in the acid soil regions. Recently released Brachiaria hybrid cv. Mulato (CIAT 36061) was identified as not only adapted to long dry season stress but also superior in its nutritional (protein) quality of green leaf forage to animals. • Arachis accessions with superior tolerance to drought identified. Field testing of 13 accessions of A. pintoi did not result in identification of specific shoot attributes that are related to superior drought adaptation. Further research work is needed to evaluate shoot and root attributes under controlled conditions in the glasshouse. • Advanced in the development of an improved screening method to evaluate drought tolerance in Brachiaria. Field screening of Brachiaria accessions at Atenas, Costa Rica were not very successful in identifying specific shoot attributes that contribute to superior adaptation to drought. Further research work is needed to evaluate shoot and root attributes under controlled conditions in the glasshouse to develop improved screening methods to evaluate drought adaptation of Brachiaria accessions and hybrids. 3.2.1 Determination of the genotypic variation in dry season tolerance in Brachiaria accessions and genetic recombinants in the Llanos of Colombia Contributors: I.M. Rao, J. W. Miles, C. Plazas, J. Ricaurte and R. García (CIAT) Rationale Quantity and quality of dry season feed is a major limitation to livestock productivity in subhumid regions of tropical America. A field study is in progress at Matazul Farm in the Llanos of Colombia. The main objective was to evaluate genotypic differences in dry season (4 months of moderate drought stress) tolerance of most promising genetic recombinants of Brachiaria. Last year, we showed that the superior performance of the Brachiaria hybrid, FM9503-S046-024, which maintained greater proportion of green leaves during dry season during the first year of establishment, was associated with lower levels of K and N content in green leaves. This year, we continued our efforts to monitor the dry season performance into second year after establishment. Materials and Methods A field trial was established on a sandy loam oxisol at Matazul farm in the Llanos of Colombia in July, 1999. The trial comprises 12 entries, including six natural accessions (four parents) and six genetic recombinants of Brachiaria. Among the germplasm accessions, B. brizantha (CIAT 26110) was identified from previous work in Atenas, Costa Rica as an outstanding genotype for tolerance to long dry season (up to 6 months). The trial was planted as a randomized block in split-plot arrangement with two levels of initial fertilizer application (low: kg/ha of 20P, 20K, 33Ca, 14 Mg, 10S; and high: 80N, 50P, 100K, 66Ca, 28Mg, 20S and micronutrients) as main plots and genotypes as sub-plots. Live and dead forage yield, shoot nutrient composition, and shoot nutrient uptake were measured at the end of the dry season (20 months after establishment; March 2001). 132 Results and Discussion Initial application of high amounts of fertilizer (at the time of establishment) did not improve forage yield of most of the genotypes compared with low fertilizer application (Table 1). This indicates very little residual effects of initial application into the second year. At 20 months after establishment (4 months after dry sseason), live forage yield with low fertilizer application ranged from 0.14 to 1.48 t/ha and the highest value of forage yield was observed with one spittlebug resistant genetic recombinant, FM9503- S046-024 (Table 75). Table 75. Genotypic variation as influenced by fertilizer application in live shoot biomass, dead shoot biomass and total forage yield of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 20 months after establishment (at the end of the dry season - March 2001). LSD values are at the 0.05 probability level. Live shoot biomass Dead shoot biomass Total forage yield Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (kg/ha) Recombinants: BR97NO-0082 475 407 204 286 679 693 BR97NO-0383 502 449 223 310 725 759 BR97NO-0405 478 690 287 380 765 1070 cv. Mulato (CIAT 36061) 985 702 611 377 1596 1079 CIAT 36062 440 609 639 476 809 1085 FM9503-5046-024 1485 1106 626 510 2111 1616 Parents: CIAT 606 548 541 560 590 1108 1131 CIAT 6294 785 727 385 382 1170 1109 BRUZ/44-02 141 196 85 64 226 260 CIAT 26646 835 1077 792 1115 1627 2192 Accessions: CIAT 26110 1008 1266 407 732 1415 1998 CIAT 26318 1074 806 754 702 1828 1508 Mean 730 715 442 494 1172 1208 LSD (P=0.05) 552 452 379 438 821 786 As expected, the performance of one of the parents, BRUZ/44-02 was very poor compared with other parents and genetic recombinants. Among the four parents, CIAT 26646 performed better but it had greater amounts of dead biomass than the other test materials.The superior performance of the hybrid, FM9503-S046-024 was mainly attributed to its ability to produce green leaf biomass during dry season (Table 76). But this hybrid produced less stem biomass than another hybrid, FM 9201-1873. Among the parents, CIAT 26646 showed greater leaf and stem biomass (Table 76). Results on leaf and stem N content indicated that BRUZ/44-02 had greater amount of N per unit leaf dry weight but its ability to acquire N (shoot N uptake) with low fertilizer application was lowest compared with other parents and genetic recombinants (Table 77). Shoot N uptake with low fertilizer application was greater for two accessions (CIAT 26110 and 26318), one parent (CIAT 26646) and one genetic recombinant (FM9503-S046-024). This genetic recombinant was also outstanding in its ability to acquire greater amounts of P, K, Ca and Mg from low fertilizer 133 application when compared with parents, accessions and other genetic recombinants (Tables 4 and 5). Among the parents, CIAT 26646 and CIAT 6294 were superior in P, K, Ca and Mg acquisition from low fertilizer application. Table 76. Genotypic variation as influenced by fertilizer application in leaf biomass, stem biomass and leaf to stem ratio of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 20 months after establishment (at the end of the dry season - March 2001). LSD values are at the 0.05 probability level. Leaf biomass Stem biomass Genotype Low Fertilizer High Low Fertilizer Fertilizer High Fertilizer Kg/ha Recombinants: BR97NO-0082 436 383 39 24 BR97NO-0383 431 381 71 68 BR97NO-0405 389 561 89 129 cv. Mulato (CIAT 36061) 493 605 492 97 CIAT 36062 402 563 38 46 FM9503-5046-024 1320 1044 165 62 Parents: CIAT 606 366 375 182 166 CIAT 6294 672 648 113 79 BRUZ/44-02 115 176 26 20 CIAT 26646 595 648 240 429 Accessions: CIAT 26110 827 985 181 281 CIAT 26318 640 581 434 225 Mean 557 579 173 136 LSD (P=0.05) 400 342 308 154 Table 77. Genotypic variation as influenced by fertilizer application in leaf N content, stem N content and shoot N uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 20 months after establishment (at the end of the dry season - March 2001). LSD values are at the 0.05 probability level. Leaf N content Stem N content Shoot N uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer % % (kg/ha) Recombinants BR97NO-0082 1.360 1.140 ND ND ND ND BR97NO-0383 1.230 1.190 ND ND ND ND BR97NO-0405 0.910 0.970 1.11 0.83 3.72 6.31 cv. Mulato (CIAT 36061) 1.260 1.040 0.48 1.03 6.39 6.85 CIAT 36062 1.200 1.120 ND ND ND ND FM9503-5046-024 1.070 0.910 0.99 0.59 15.35 10.01 Parents: CIAT 606 1.360 1.180 0.84 0.91 6.21 5.10 CIAT 6294 0.990 0.900 1.05 0.67 6.99 6.06 BRUZ/44-02 2.240 1.900 ND ND ND ND CIAT 26646 1.060 0.960 0.68 0.67 8.05 8.72 Accessions: CIAT 26110 1.040 0.870 0.90 0.84 10.12 10.53 CIAT 26318 0.970 1.110 0.65 0.81 9.08 7.41 Mean 1.160 1.070 0.8 0.8 7.21 6.78 LSD (P=0.05) 0.451 0.481 NS 0.36 6.11 4.58 ND = not determined due to small size of the sample; NS = not significant. 134 Table 78. Genotypic variation as influenced by fertilizer application in leaf P content, stem P content and shoot P uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 20 months after establishment (at the end of the dry season. Leaf P content Stem P content Shoot P uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer % % (kg/ha) Recombinants BR97NO-0082 0.116 0.126 ND ND ND ND BR97NO-0383 0.088 0.108 ND ND ND ND BR97NO-0405 0.101 0.116 0.124 0.130 0.42 0.78 cv. Mulato (CIAT 36061) 0.108 0.109 0.090 0.103 0.62 0.71 CIAT 36062 0.132 0.163 ND 0.152 ND 0.97 FM9503-5046-024 0.116 0.119 0.136 0.184 1.75 1.33 Parents: CIAT 606 0.118 0.137 0.117 0.172 0.63 0.76 CIAT 6294 0.113 0.109 0.165 0.099 0.87 0.74 BRUZ/44-02 0.092 0.143 ND ND ND ND CIAT 26646 0.104 0.127 0.084 0.106 0.82 1.24 Accessions: CIAT 26110 0.106 0.106 0.128 0.165 1.06 1.47 CIAT 26318 0.095 0.114 0.088 0.107 0.97 0.89 Mean 0.108 0.123 0.112 0.130 0.76 0.85 LSD (P=0.05) NS 0.039 0.040 0.053 0.68 0.63 ND = not determined due to small size of the sample; NS = not significant. Table 79. Genotypic variation as influenced by fertilizer application in shoot K uptake, shoot Ca uptake and shoot Mg uptake of genetic recombinants, parents and other germplasm accessions of Brachiaria grown in a sandy loam oxisol at Matazul, Colombia. Plant attributes were measured at 20 months after establishment (at the end of the dry season - March 2001). LSD values are at the 0.05 probability level. Shoot K uptake Shoot Ca uptake Shoot Mg uptake Genotype Low High Low High Low High Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer Fertilizer (kg/ha) Recombinants: BR97NO-0082 7.62 7.24 0.90 1.23 0.83 1.15 BR97NO-0383 7.59 6.17 1.01 0.95 1.17 0.94 BR97NO-0405 7.00 11.86 1.08 1.93 1.08 1.99 cv. Mulato (CIAT 36061) 6.94 9.38 1.26 1.59 1.41 1.81 CIAT 36062 7.96 9.35 0.91 1.53 0.87 2.09 FM9503-5046-024 23.58 13.32 3.70 3.55 3.69 3.63 Parents: CIAT 606 9.55 8.16 1.01 1.04 1.30 1.62 CIAT 6294 10.48 9.53 1.47 1.49 1.75 1.82 BRUZ/44-02 2.49 4.20 0.30 0.75 0.31 0.85 CIAT 26646 9.50 13.95 1.28 1.84 1.62 2.88 Accessions: CIAT 26110 14.10 15.54 1.73 2.51 1.94 3.60 CIAT 26318 12.05 8.95 1.67 1.40 2.10 2.12 Mean 10.40 9.92 1.43 1.67 1.59 2.07 LSD (P=0.05) 8.25 6.14 1.70 1.22 1.73 1.60 135 Correlation analysis between green leaf biomass produced in the dry season and other shoot attributes indicated that superior performance with low fertilizer application was associated with lower level of N in green leaves (Table 80). Significant negative association was also observed between green leaf biomass and lower level of K and N in green leaves with high fertilizer application. This observation indicates that genotypes that are efficient in uitilization of N for the production of green forage is an important mechanism for superior performance with low fertilizer application in the dry season. Conclusions Results from this field study indicated that the superior performance of the Brachiaria hybrid, FM9503- S046-024 which maintained greater proportion of green leaves during moderate dry season in the llanos of Colombia, was associated with lower levels of K and N content in green leaves. Table 80. Correlation coefficients (r) between green leaf biomass (t/ha) and other shoot traits of Brachiaria genotypes grown with low or high fertilizer application in a sandy loam oxisol in Matazul. Shoot traits Low fertilizer High fertilizer Live forage yield (t/ha) 0.87*** 0.72*** Total forage yield (t/ha) 0.81*** 0.81*** Dead biomass (t/ha) 0.54*** 0.51*** Stem biomass (t/ha) 0.20 0.45** Leaf N content (%) -0.33* -0.45** Leaf P content (%) 0.05 -0.10 Leaf K content (%) -0.15 -0.54*** Leaf Ca content (%) -0.16 0.06 Stem N content (%) 0.20 -0.11 *,**,*** Significant at the 0.05, 0.01 and 0.001 probability levels, respectively. 3.2.2 Determination of the genotypic variation in dry season tolerance in Brachiaria and Arachis in Costa Rica Contributors: I.M.Rao, P. J. Argel, J. Ricaurte and R. García (CIAT) Rationale Field studies were continued at Atenas, Costa Rica. The main objective was to evaluate genotypic differences in dry season (6 months) tolerance among 16 accessions of Brachiaria species and 13 accessions of Arachis species. We continued our work to test further the hypothesis that tolerance to dry season is greater in genotypes that accumulate greater amounts of total nonstructural carbohydrates (TNC) combined with less amounts of minerals (ash content) per unit dry weight of leaves and stems. The use of shoot attributes such as ash content, Ca content and nonstructural carbohydrate levels as selection criteria for dry season tolerance in Brachiaria is being tested further using green leaves developed during dry season compared to dry leaf and stem tissue. 136 Materials and Methods Trial 1 included 16 genotypes (15 accessions and 1 hybrid) of Brachiaria species and trial 2 included 13 accessions of Arachis species selected from agronomic evaluation of the germplasm. Atenas site provided excellent field conditions to evaluate the impact of long dry season (5 months) while keeping nutrient supply in soil adequate for growth. Forage yield, nutrient composition, and nonstructural carbohydrates and ash content in green leaves, dry leaves and stem tissue were measured. Results and Discussion Trial 1 - Forage yield among Brachiaria species during dry season ranged from 2988 to 9988 kg/ha and the greatest forage yield was observed with B. brizantha CIAT 26646 (Table 81). The superior performance of this accession at Atenas site is consistent with its outstanding performance at Matazul site in the Llanos of Colombia (see above Activity 3.2.1). B. brizantha CIAT 26110, which maintained greater proportion of green leaves during dry season (visual observation) maintained greater amountn of N (crude protein) and TNC in green leaves while its ash (mineral) content was markedly lower than most accessions. This observation indicates that this accession combines drought adaptation with greater nutritional value of the green forage. One of the Brachiaria hybrids tested (CIAT 36061) was particularly outstanding in its N status of the green leaves. It apperas that this hybrid also combined adaptation to drought with greater nutritional value of the green forage. Among the 16 genotypes B. brizantha CIAT 667 was outstanding in maintaining greater amounts of nonstructural carbohydrates in green leaves. Table 81. Genotypic variation in forage yield, green leaf nutrient composition, ash content and total nonstructural carbohydrates (TNC) of 16 genotypes of Brachiaria species grown during dry season at Atenas, Costa Rica. Genotype Forage yield Green leaf composition (CIAT number) (kg/ha) C N P K Ca Mg Ash TNC (%) (mg/kg) B. brizantha (26646) 9988 25.3 0.45 0.074 1.345 0.546 0.648 9.0 156 B. brizantha (16305) 8828 27.8 0.67 0.057 1.468 0.418 0.513 8.3 141 B. brizantha (16322) 8132 26.5 0.56 0.050 1.150 0.566 0.721 11.4 174 B. brizantha (16319) 8068 27.0 0.39 0.038 1.379 0.337 0.517 6.8 180 B. brizantha (26110) 7028 27.1 0.71 0.049 1.472 0.409 0.295 7.5 178 cv. Mulato (CIAT 36061) 6692 28.0 0.81 0.062 1.334 0.701 0.755 11.4 161 B. brizantha (16300) 6640 26.9 0.47 0.045 1.139 0.495 0.623 7.1 158 B. brizantha (16467) 6628 27.2 0.57 0.076 1.860 0.606 0.519 9.6 146 B. brizantha (667) 6492 25.5 0.69 0.079 1.586 0.506 0.437 8.4 263 B. brizantha (16168) 5548 26.9 0.49 0.038 1.445 0.656 0.622 10.3 93 B. brizantha (16549) 5372 26.5 0.50 0.050 1.151 0.569 0.525 7.8 164 B. brizantha (16289) 5308 27.1 0.51 0.050 1.494 0.564 0.477 9.7 199 B. briznahta (16488) 5080 26.4 0.50 0.069 1.161 0.564 0.587 10.8 121 B. brizantha (16135) 4548 26.4 0.73 0.077 1.054 0.651 0.828 8.8 154 B. decumbens (16497) 3492 26.2 0.68 0.063 1.205 0.546 0.587 7.7 125 B. brizantha (6387) 2988 25.7 0.76 0.101 1.126 0.881 0.742 10.7 184 Mean 6302 26.7 0.59 0.061 1.336 0.563 0.587 9.1 162 Results on composition of dry leaves in terms of nutrients and TNC indicated that B. brizantha CIAT 667 is also outstanding in maintaining greater levels of TNC and nutrients (Table 82). Among the 16 genotypes tested. B. brizantha CIAT 16300 showed the lowest amount of ash content of both green leaves and dry leaves indicating that this genotype had greater nutrient use efficiency to produce forage 137 during dry season. Genotypic variation in nutrient and TNC composition of stems indicated that the hybrid CIAT 36061 is outstanding in maintaining greater composition of N and TNC than other genotypes. These results indicate that the hybrid is not only productive during dry season but also nutritive to the animals (Table 82). Table 82. Genotypic variation in dry leaf nutrient composition, ash content and total nonstructural carbohydrates (TNC) of 16 genotypes of Brachiaria species grown during dry season at Atenas, Costa Rica. Genotype Dry leaf composition (CIAT number) C N P K Ca Mg Ash TNC (%) (mg/g) B. brizantha (26646) 25.4 0.32 0.073 0.873 0.537 0.629 9.5 140 B. brizantha (16305) 27.0 0.42 0.028 0.769 0.265 0.302 8.9 129 B. brizantha (16322) 26.5 0.36 0.048 0.868 0.561 0.565 11.0 134 B. brizantha (16319) 26.5 0.18 0.027 1.099 0.379 0.677 8.1 137 B. brizantha (26110) 26.5 0.51 0.041 1.211 0.460 0.361 8.9 154 cv. Mulato (CIAT 36061) 26.7 0.41 0.041 0.630 0.624 0.609 11.8 152 B. brizantha (16300) 28.6 0.33 0.054 0.780 0.568 0.631 7.7 247 B. brizantha (16467) 26.9 0.23 0.059 1.146 0.479 0.438 9.7 159 B. brizantha (667) 26.4 0.34 0.055 0.972 0.545 0.468 10.6 258 B. brizantha (16168) 28.7 0.43 0.044 1.098 0.547 0.533 10.1 201 B. brizantha (16549) 25.2 0.35 0.036 0.809 0.570 0.548 8.9 146 B. brizantha (16289) 27.9 0.48 0.054 1.027 0.479 0.398 9.0 139 B. brizanhta (16488) 27.0 0.38 0.065 1.012 0.435 0.540 10.9 143 B. brizantha (16135) 26.8 0.68 0.096 1.025 0.514 0.717 10.3 134 B. decumbens (16497) 26.5 0.58 0.074 0.870 0.573 0.560 8.6 90 B. brizantha (6387) 28.1 0.70 0.095 0.823 0.699 0.533 10.6 151 Mean 26.9 0.42 0.056 0.938 0.515 0.532 9.7 157 Trial 2- Although forage yield data were not available, among the 13 accessions of Arachis pintoi tested, three accessions, CIAT 17434, 22159 and 22161 maintained greater concentration of N in green leaf tissue (Table 83). One of the accessions, A. pintoi CIAT 22160, which was selected as dry season tolerant accession from field evaluation in cerrados of Brazil showed level of K in green leaf tissue. It also showed greater levels of Ca, Mg and TNC in green leaf tissue. Results on genotypic variation in dry leaf nutrient composition, ash content and TNC content showed that CIAT 22160 had lower levels of TNC indicating that it may have greater ability to mobilize photosynthates (TNC) from older dry leaves to young green leaves (Table 84). This may be an important mechanism in the shoots in addition to its better rooting ability to avoid drought stress. Further research is needed to test this accession compared with the commercial check, CIAT 17434 under glasshouse conditions to compare root and shoot attributes with drought stress. Conclusions Results from the above 2 trials did not provide any clear evidence that using green leaf nutrient status, ash content or TNC one good indicator of drought adaptation of Brachiaria and Arachis gentoypes. 138 Table 83. Genotypic variation in stem nutrient composition, ash content and total nonstructural carbohydrates (TNC) of 16 genotypes of Brachiaria species grown during dry season at Atenas, Costa Rica. Genotype Stem composition (CIAT number) C N P K Ca Mg Ash TNC (%) (mg/g) B. brizantha (26646) 27.3 0.11 0.040 0.374 0.169 0.334 3.9 144 B. brizantha (16305) 24.9 0.19 0.047 0.752 0.117 0.233 4.8 165 B. brizantha (16322) 26.1 0.19 0.042 0.582 0.148 0.202 5.8 116 B. brizantha (16319) 26.6 0.21 0.040 0.721 0.144 0.362 4.4 174 B. brizantha (26110) 25.4 0.30 0.046 0.723 0.151 0.157 4.4 99 cv. Mulato (CIAT 36061) 26.4 0.50 0.039 0.528 0.208 0.292 4.9 267 B. brizantha (16300) 27.4 0.25 0.050 0.561 0.164 0.374 4.6 153 B. brizantha (16467) 27.5 0.13 0.044 0.973 0.161 0.276 5.0 142 B. brizantha (667) 26.4 0.14 0.056 0.999 0.170 0.208 6.5 153 B. brizantha (16168) 27.6 0.07 0.025 0.513 0.110 0.179 4.4 152 B. brizantha (16549) 27.2 0.19 0.035 0.417 0.185 0.218 3.6 152 B. brizantha (16289) 24.8 0.11 0.024 0.863 0.141 0.193 5.6 150 B. brizanhta (16488) 26.3 0.11 0.042 0.600 0.151 0.320 6.4 133 B. brizantha (16135) 26.7 0.19 0.044 0.618 0.133 0.341 5.6 95 B. decumbens (16497) 26.0 0.26 0.057 0.626 0.145 0.200 4.7 116 B. brizantha (6387) 26.5 0.20 0.071 0.608 0.149 0.186 5.6 214 Mean 26.4 0.20 0.044 0.654 0.153 0.255 5.0 152 Table 84. Genotypic variation in green leaf nutrient composition, ash content and total non- structural carbohydrates (TNC) of 13 accessions of Arachis pintoi grown during dry season at Atenas, Costa Rica. Genotype (CIAT Green leaf composition number) C N P K Ca Mg Ash TNC (%) (mg/kg) A.pintoi (17434) 28.8 2.67 0.119 0.959 2.470 0.861 8.9 104 A.pintoi (18744) 26.2 2.62 0.103 0.785 2.423 0.981 9.1 69 A. pintoi (22148) 25.4 2.21 0.069 0.834 2.363 0.534 8.3 139 A. pintoi (22149) 26.3 2.48 0.091 0.929 2.371 0.501 8.8 127 A. pintoi (22150) 26.3 2.16 0.071 1.017 2.307 0.579 8.7 143 A. pintoi (22151) 26.3 2.31 0.076 0.571 2.590 0.710 8.9 64 A. pintoi (22155) 26.0 2.26 0.076 1.284 2.280 0.502 8.8 138 A. pintoi (22156) - - - - - - - - A. pintoi (22157) 26.5 2.62 0.105 1.175 2.193 0.612 9.0 122 A. pintoi (22158) 26.7 2.47 0.086 0.956 2.403 0.765 9.5 143 A. pintoi (22159) 26.6 2.67 0.094 0.811 2.398 0.857 9.4 73 A. pintoi (22160) 27.2 2.35 0.077 0.490 2.511 1.004 9.5 117 A. pintoi (22161) 29.9 2.67 0.112 0.952 2.030 0.933 8.5 95 Mean 26.9 2.46 0.090 0.897 2.362 0.680 9.0 111 139 Table 85. Genotypic variation in dry leaf nutrient composition, ash content and total nonstructural carbohydrates (TNC) of 13 accessions of Arachis pintoi grown during dry season at Atenas, Costa Rica. Genotype Dry leaf composition TNC (CIAT number) C N P K Ca Mg Ash (mg/g) A. pintoi (17434) 26.2 2.03 0.089 0.681 2.487 0.861 9.2 63 A. pintoi (18744) 28.6 2.15 0.092 0.654 2.886 0.981 10.3 42 A. pintoi (22148) 26.7 1.87 0.060 0.607 2.304 0.534 8.3 90 A. pintoi (22149) 26.3 2.21 0.081 0.544 3.096 0.501 8.2 80 A. pintoi (22150) 27.3 1.79 0.057 0.793 2.337 0.579 9.0 64 A. pintoi (22151) 26.7 1.59 0.045 0.377 3.017 0.710 10.3 61 A. pintoi (22155) 27.1 1.82 0.056 0.968 2.508 0.502 9.2 91 A. pintoi (22156) 26.7 1.90 0.055 0.789 2.524 - 9.1 53 A. pintoi (22157) 24.8 1.97 0.076 0.747 2.760 0.612 9.4 109 A. pintoi (22158) 27.1 2.11 0.088 0.685 2.798 0.765 9.8 82 A. pintoi (22159) 27.3 2.01 0.066 0.585 2.527 0.857 9.5 36 A. pintoi (22160) 28.5 1.99 0.069 0.355 2.598 1.004 9.3 52 A. pintoi (22161) 26.0 2.05 0.074 0.970 2.341 0.933 9.0 134 Mean 26.9 1.96 0.070 0.673 2.629 0.737 9.3 73 Activity 3.3 Shrub legumes with adaptation to drought and cool temperatures Highlights • Found significant differences within and between a collection of Cratylia argentea accessions and Leucaena spp. in quality attributes. • Accesssions of Cratylia argentea with superior performance than C. argentea cv. Veraniega identified • Developed map of potential distribution of Flemingia macrophylla in tropical Asia and found genetic variability for forage quality parameters in this legume species. Progress towards achieving milestones • List of new accessions of Cratylia argentea and Leucaena species with known forage value Our results show considerable variability in growth habit, DM yields, and quality parameters in accessions of C. argentea, which open the opportunity for the selection of new cultivars in the near future. Accessions CIAT 18674, 22375, 22406, 22408 and 22409 had higher dry matter yields than CIAT 18516/18668 (cv. Veraniega) in dry and wet seasons. The new accessions have also showed dry tolerance and good re-growth during prolonged dry seasons, which is one of the outstanding characteristics that makes C. argentea a valuable forage for dual purpose cattle farms. The legume Leucaena leucocephala is very well known for its high forage value. However, a great diversity exists within this genus that has not been fully characterized. For instance, our results showed that L. macrophylla susp. nelsonii OFI 47/85, species commonly found along the coasts of Oxaca and Guerrero in Mexico, showed high CP content (28.3%), and acceptable IVDMD (62.2%), indicating that it deserves to be evaluated with animals in futures studies. • List of Flemingia macrophylla accessions characterized for yield and quality Results indicate that several accessions have superior dry matter yield and better digestibility than the control CIAT 17403. The most promising accession (CIAT 21090) will be multiplied for further 140 testing. Studies to better understand the difference in digestibility among Flemingia accessions were initiated and these studies will be complemented with palatability trials. 3.3.1 Characterization of a core collection of Cratylia argentea and Leucaena sp. in a subhumid environment of Costa Rica Contributors: Diego Bolaños Crespo, Mayra Montiel L., Miguel Vallejo Solís y Carlos Jiménez Crespo (U. de Costa Rica); Guillermo Pérez y P. J. Argel (CIAT) Rationale One of the limitations of Leucaena leucocephala is its limited adaptation to acid soils and to pest such as psyllid. On the other hand, C. argentea, a shrub native to the South American tropics, has shown good adaptation to acid soils, and excellent tolerance to prolonged dry periods. Thus, we were interested in characterizing core collections of C. argentea and Leucaena for yield and quality attributes in a site characterized by having a long dry season and acid soils of medium fertility. Materials and Methods Leaves and young stems (edible forage) were harvested from 18 lines of Leucaena spp. and 30 lines of Cratylia argentea planted in Atenas, Costa Rica, that had been under cutting evaluation for 2 years. The site is located in a subhumid environment with a total annual rainfall of 1600 mm, and 5 to 6 months dry from December to May. The soils are Inceptisol of medium fertility with pH 5.0, and low P and low aluminum content. The samples were dried in an air forced oven set at 60 °C for 72 hours, and then ground to less than 5-mm particles. Dry matter (DM) and protein content (CP) were determined using standard AOAC procedures. Neutral-detergent fibre (NDF) and acid detergent fibre (ADF) were determined according to Van Soest and Roberston (1979), while invitro dry matter digestibility (IVDMD) was determined as described by Tilley and Terry (1963). Results and Discussion Results on quality of C. argentea and Leucaena spp. are presented in Table 86 and Table 87, respectively. The mean CP was higher in Leucaena sp. (mean of 23.3%) compared to Cratylia (mean of 17.7%). L. macrophylla subsp. nelsonii OFI 47/85 had the highest value of CP (28.3%), while L. pulverulenta OFI 83/87 had the lowest (17.2%). In general, variation of CP content within Cratylia argentea accessions was less than in the Leucaena species, which indicates more genetic variability within the latter group. On the other hand, C. argentea CIAT 22382 had the highest CP content (19.4%), and C. argentea CIAT 22389 the lowest CP value (15.1%). The NDF contents for both Cratylia and Leucaena accessions are within the range reported in the literature for tropical tree legumes. The NDF fraction varied from 43-80% for the Leucaena species, and from 62-77% for C. argentea. Considerable interspecific variation in NDF content existed within the species of Leucaena, evaluated with L. macrophylla subsp. nelsonii OFI 47/85 showing the highest content of NDF (80%) while L. diversifolia subsp. diversifolia OFI 83/92 the lowest (43%). There was little variability of NDF values within the Cratylia group, indicating a good degree of genetic uniformity for the 30 accessions evaluated. 141 Table 86. Quality components of Cratylia argentea accesions established in the subhumid enviroment of Atenas, Costa Rica. CIAT No. DM (%) CP (%) NDF (%) ADF (%) IVDMD* (%) 22374 34.2 15.8 68.6 52.7 54.0 22386 30.8 16.8 69.9 50.2 60.9 22379 31.6 16.6 69.5 50.9 50.5 22382 33.0 19.4 67.3 46.4 64.9 22381 31.0 18.8 67.7 51.5 56.5 22375 32.8 15.7 66.5 49.9 54.1 22389 34.0 15.1 68.7 48.6 55.7 (BRA 000621) 42.1 18.2 71.1 57.0 - 22391 33.3 16.6 61.6 44.9 51.3 (BRA 000876) 31.6 19.3 62.3 45.3 61.6 22393 31.9 18.4 63.6 44.2 65.0 22378 32.3 18.5 68.1 46.4 58.8 (Yacapani) 32.9 18.6 64.4 43.2 55.8 22380 34.3 17.5 68.9 50.6 53.1 22385 37.2 17.1 65.4 45.6 54.8 22383 32.2 16.2 69.7 47.8 53.7 22384 34.5 16.1 65.8 48.2 57.2 22395 33.8 19.0 67.7 48.4 55.1 22373 32.1 18.6 67.6 52.2 57.1 22390 32.2 18.7 66.6 47.7 50.3 22392 31.2 19.0 64.2 44.8 57.3 22387 32.0 17.5 66.4 49.7 - 22394 34.2 18.4 67.9 48.7 60.3 (BRA 000884) 32.8 19.3 65.9 42.8 62.5 22388 31.2 18.2 68.4 48.4 58.9 (BRA 000604) 37.3 17.8 65.3 46.5 55.4 (BRA 000841) 34.5 18.4 65.6 47.1 55.9 22396 32.5 19.0 64.6 47.1 47.8 22377 33.8 17.9 67.3 54.4 48.1 22376 32.0 17.3 76.7 48.6 57.4 Mean 33.2 17.7 69.1 48.1 55.9 Sd 2.3 1.2 2.9 3.2 4.4 *Quality components of eatable forage (leaves and young stems) 8 weeks old The mean level of ADF was slightly higher in Cratylia than in the Leucaena accessions, but within the range expected for tropical legumes. On the other hand, DM digestibility was higher, and more variable in Leucaena (range from 50-84%), compared to Cratylia (range from 48-65%). Forage quality parameters of C. argentea cv. Veraniega (CIAT 18516/18668) at 90 days of re-growth, are within the range observed in the core collection of C. argentea accessions evaluated. On the other hand, L. leucocephala subsp. glabrata cv. Taramba en Australia (OFI 34/92), showed high CP (33%) and IVDMD (65%), indicating a high potential feed value; however, this line is susceptible to the psyllid, which may limit its commercial use in sites with high incidence of the insect. 142 Table 87. Quality components of Leucaena species established in the subhumid environment of Atenas, Costa Rica. Species ID No. DM CP NDF ADF IVDMD* (OFI) (%) (%) (%) (%) (%) L. trichanda 53/88 29.2 24.9 50.6 46.6 51.8 L collinsii 52/88 29.5 26.6 45.1 35.5 84.0 L. leucocephala subsp. glabrata 34/92 32.6 23.8 65.7 40.3 64.6 L. pallida 14.96 30.6 22.9 70.6 53.5 53.4 L. hybrid 1/95 31.8 21.3 63.3 42.3 - L. macrophylla subsp. nelsonii 47/85 43.4 28.3 80.2 45.4 62.2 L. leucocephala CIAT 17263 31.4| 24.3 46.9 41.6 71.5 Leucaena hybrid 52/87 32.3 25.0 52.3 41.3 - L. salvadorensis 17/86 35.9 22.4 49.7 38.5 69.9 L. lanceolata 43/85 33.8 24.2 46.4 41.1 73.3 L. diversifolia subsp. diversifolia 83/92 30.1 26.1 43.4 38.0 55.7 L. pallida 79/92 31.4 24.4 55.2 49.6 50.1 L. esculenta subsp. esculenta 47/87 33.6 20.7 68.0 40.1 60.6 L. pulverulenta 83/87 38.0 17.2 62.7 47.4 73.1 L. collinsii subsp. zacapana 56/88 30.5 21.7 52.3 46.3 70.9 L. lempirana 6/91 31.3 24.5 47.6 36.0 80.7 L. shannonii subsp. magnifica 19/84 31.2 23.5 51.0 39.9 75.7 L. trichodes 61/88 32.3 22.7 55.9 45.2 66.6 Mean 32.4 23.3 54.3 42.2 64.9 Sd 3.4 2.2 10.2 4.8 10.2 *Quality components of eatable forage (leaves and young stems) 8 weeks old 3.3.2 Genetic diversity in the multipurpose shrub legumes Flemingia macrophylla and Cratylia argentea Contributors: M. Andersson (University of Hohenheim), M. Peters, J. Tohme (CIAT), R. Schultze-Kraft (University of Hohenheim), and L.H. Franco (CIAT) CIAT projects: SB-2 Rationale Work on shrub legumes in CIAT, emphasizes the development of species to be utilized as feed supplement during extended dry periods. Tropical shrub legumes of high quality for better soils are readily available, but germplasm with similar characteristics adapted to acid, infertile soils is scarce. Shrub legume species, such as Flemingia macrophylla and Cratylia argentea are well adapted to low fertility soils and to prolonged drought, respectively. Thus, work on these genera is of high priority in CIAT's Forage Project. In order to define the extent of genetic diversity within ex-situ collections of F. macrophylla and C. argentea we initiated a project with three main objectives. (a) to identify new, superior forage genotypes based on conventional germplasm characterization/evaluation procedures (morphological and agronomic traits, forage quality parameters, including IVDMD and tannin contents), and (b) to optimize the use and management, including conservation, of the collections. To accomplish these objectives, different approaches are being used with a core collection: (a) genetic diversity assessment by a germplasm origin information; and (b) molecular markers (AFLPs). This information should also be useful to define future collection needs in terms of geographical focus. 143 Materials and Methods Agronomic characterization and evaluation: Spaced-plants of Cratylia argentea (39 accessions) and Flemingia macrophylla (73 accessions) were established in Quilichao in March 1999 (Photo 13) and March 2000 (Photo 14), respectively. Additionally two replications were sown for morphological characterization and for seed production. The following parameters are being measured: vigor, height and diameter, regrowth, seasonal dry matter yield during incidence of diseases, pests and mineral deficiencies. Photo 13. Cratylia argentea at Quilichao Photo 14. Flemingia macrophylla accession at Quilichao 144 For the morphological evaluation, qualitative and quantitative parameters are recorded, such as days to first flower, days to first seed, flower color, flowers per inflorescence, flowering intensity, pod pubescence, seeds per pod, seed color, branching capacity, leaf length and width, peduncle length, etc. To assess nutritive value, we are measuring crude protein (CP) and in vitro dry-matter digestibility (IVDMD) in leaf samples of the two collections. For F. macrophylla, a more detailed chemical analysis will be conducted on a representative subset which will include accessions with high intermediate and low CP and IVDMD. Other chemical analysis in selected accessions of F. macrophylla will include NDF, ADF, condensed tannins calcium, and phosphorus. Analysis of available origin information: Based on geographical information on the site of origin of accessions, a core collection will be created under the assumption, that geographic distances and environmental differences are related to genetic diversity. The analysis will be conducted with FloraMapTM, a GIS tool developed by CIAT Genetic analysis by molecular markers (AFLPs): Genetic variability will also be assessed through AFLP molecular markers. Based on the results of molecular markers group of accessions will be formed, using multivariate statistic tools. Data analysis and synthesis: Individual and combined analyses of all data generated will be performed, including the use of GIS tools and multivariate statistics. In the analysis of each of the different approaches (agronomic characterization, origin information, molecular marker analysis), Principal Component Analysis and Cluster Analysis will be utilized to assist in the formation of core collections. Correlation between the different approaches and clusters obtained will also be determined. These results are expected to help in deciding which of the three methods or which combination is most appropriate (time and cost efficiency) to create a core collections. For example, if an agronomic evaluation of the entire collection is not feasible because of time constraints, a core collection may be created using origin information and/or molecular markers. In addition, based on similarity of molecular marker and GIS analysis, we hope to provide information that will be useful for defining future collections on areas with particularly high diversity. Accession duplicates in the world collections will also be identified. Results and Discussion Based on the origin of existing germplasm accessions (73), the potential natural distribution of F. macrophylla extends throughout vast areas of tropical Asia (Photo 15). This map, however, is not to be considered as a ‘prediction’ of the probable natural distribution of F. macrophylla but merely as an indication of conditions based on climate and latitude/longitude matching those of the germplasm collection sites. Agronomic characterization and evaluation: Results from one evaluation in the dry season and one in the rainy season show that there is considerable phenotypic and agronomic variation in the collection of Cratylia argentea evaluated (Table 88) and Flemingia macrophylla (Table 89). In the case of C. argentea mean dry matter production was 45 g/plant in the wet and 60 g/plant in the dry season, whereas IVDMD varied between 61 and 67% and CP content between 18 and 21%. 145 Photo 15. FLORAMAP analysis of potential natural distribution of Flemingia macrophylla in tropical Asia, based on passport data of the world germplasm collection maintained at CIAT. Principal component analysis performed with the agronomic data of 39 accessions of C. argentea revealed high correlations between total dry matter production, diameter, regrowth points and vigour. Cluster analysis resulted in 9 groups and 5 of the clusters contained only one accession, among them three of the most productive accessions (CIAT 18674, 22406 and 22408). Based to these initial results, we have pre-selected C. argentea accessions CIAT 18674, 22375, 22406, 22408 and 22409 given that productivity of these accessions is higher than the genotypes released in Costa Rica (an accession mix of CIAT 18516/18668) as cv. Veraniega. In the case of F. macrophylla the average dry matter production was 60 g/plant in the wet and 42 g/plant in the dry season. The most productive accessions were CPI 104890, CIAT 21090, 21241, 21529 and 21580 with a total dry matter production >100 g/plant. We also found high variability in IVDMD (31 to 51%) and CP (16 to 24%), which is an interesting results since one limitation of F. macrophylla is low feed value. Principal component analysis performed with the agronomic data of 73 accessions of F. macrophylla revealed high correlations between total dry matter production, plant height and diameter and vigour (>70%). Cluster analysis (UPGMA) resulted in 7 clusters and 2 of the clusters contained only one accession, among them one of the most productive accessions (CIAT 21090). Based on these results we selected the F. macrophylla accession CIAT 21090 (semi-erect type, high forage yield and quality for seed multiplication and evaluation with animals. 146 Table 88. Agronomic evaluation of a collection of Cratylia argentea in Quilichao. Preliminary data of four cuts (two in the dry season and two in the wet season). Treatment Height Diameter Regrowing Mean dry matter yields IVDMD Crude No. CIAT (cm) (cm) points Wet Dry Mean (%) protein (No.) (g/pl) (%) 18516 112 105 19 55 78 66 65.0 20.7 18667 112 101 18 45 68 56 64.6 20.4 18668 106 110 17 48 68 58 65.2 19.9 18671 111 106 20 54 55 54 64.3 18.3 18672 96 83 13 34 39 37 62.1 20.1 18674 118 122 23 91 109 100 63.9 20.0 18675 112 97 15 47 63 55 63.3 19.0 18676 105 93 14 46 50 48 61.2 19.7 18957 111 102 16 50 76 63 62.5 20.1 22373 109 93 15 38 57 48 64.4 20.2 22374 116 102 17 55 71 63 66.4 19.6 22375 125 98 16 59 76 68 67.0 21.2 22376 95 70 11 23 36 29 64.1 19.6 22378 103 81 12 34 39 36 61.7 19.8 22379 111 89 16 47 65 56 63.5 19.6 22380 107 90 11 31 43 37 61.3 20.4 22381 105 85 11 34 46 40 64.0 19.1 22382 110 92 12 41 62 52 64.2 20.4 22383 99 90 13 34 43 39 62.6 18.6 22384 113 91 9 43 47 45 64.5 18.9 22386 111 86 12 39 47 43 64.7 18.6 22387 111 90 12 41 57 49 62.5 19.1 22390 99 92 13 45 47 46 64.8 18.5 22391 108 96 15 44 62 53 63.4 18.9 22392 114 83 13 33 53 43 63.2 21.0 22393 110 92 17 41 58 49 63.5 20.6 22394 112 88 13 33 46 40 64.0 20.5 22396 101 79 10 30 43 36 63.8 21.3 22399 102 86 13 35 42 39 66.1 19.8 22400 119 104 16 52 74 63 61.7 20.7 22404 110 97 13 42 68 55 67.0 20.9 22405 111 96 16 41 61 51 62.9 19.9 22406 113 112 20 63 86 74 62.6 21.0 22407 111 99 16 46 59 53 65.3 20.8 22408 120 109 18 69 88 79 67.2 20.1 22409 113 115 17 57 81 69 66.5 21.2 22410 116 96 14 42 60 51 64.3 19.8 22411 103 88 14 37 58 47 64.5 20.2 22412 116 90 11 42 63 52 64.9 18.7 Mean 110 95 15 45 60 52 64.1 19.9 Range 95-125 70-122 9-23 23-91 36-109 29-100 61-67 18-21 147 Table 89. Agronomic evaluation of a collection of Flemingia macrophylla in Quilichao. Preliminary data of two cuts (one in each season). Growth habit: =erect, s=semierect, p=prostrate. Treatment Height Diameter Regrowing Mean dry matter yields (g/pl) IVDMD Crude No. CIAT (cm) (cm) Points (No) Wet Dry Mean (%) Protein (%) J 001 (e) 125 85 30 102 58 80 40.1 22.3 801 (e) 125 90 29 103 62 82 36.3 22.9 7184 (e) 124 95 34 101 82 92 34.0 21.4 C 10489 (e) 108 99 34 121 79 100 33.6 22.7 I 15146 (e) 98 70 24 103 58 80 39.9 22.9 17400(s) 63 98 33 55 52 53 33.2 21.5 17403 (s) 67 96 32 68 57 62 35.8 22.2 17404 (s) 58 79 32 46 45 45 32.9 22.5 17405 (s) 65 94 36 71 67 69 36.1 21.9 17407 (s) 78 106 39 87 62 74 32.8 21.9 17409 (s) 56 109 35 87 66 77 33.0 20.2 17411 (s) 55 86 33 56 54 55 35.5 22.4 17412 (s) 73 96 39 61 63 62 38.6 20.2 17413 (s) 58 93 35 51 39 45 35.2 20.1 18048 (s) 32 43 19 12 8 10 42.8 20.4 18437 (s) 54 101 37 57 55 56 47.8 22.5 18438 (s) 58 71 31 36 22 29 51.5 23.5 18440 (s) 59 87 38 65 44 55 33.4 21.4 19453 (e) 105 78 20 65 33 49 36.0 21.6 19454 (e) 115 82 24 73 52 63 39.1 19.7 19457 (e) 116 85 25 52 64 58 33.1 21.3 19797 (s) 57 90 22 58 46 52 38.5 21.0 19798 (s) 55 95 27 61 55 58 38.3 20.9 19799 (s) 50 69 19 28 39 33 37.3 21.7 19800 (s) 65 85 29 34 48 41 32.0 20.7 19801 (s) 82 91 40 68 57 63 35.7 21.7 19824 (e) 62 93 35 54 61 58 35.6 21.3 20065 (p) 15 21 4 0 1 1 32.1 18.9 20616 (s) 67 108 34 86 54 70 32.1 22.0 20617 (s) 72 92 27 51 44 48 30.6 20.1 20618 (s) 74 95 31 57 60 58 33.2 21.6 20621 (e) 84 88 32 58 54 56 31.6 21.6 20622 (e) 146 88 30 105 76 91 42.8 22.9 20624 (s) 74 122 39 101 91 96 34.5 19.8 20625 (e) 128 86 26 105 69 87 42.5 22.8 20626 (e) 115 92 28 88 70 79 39.5 22.3 20631 (e) 121 90 25 97 75 86 41.5 20.9 20744 (e) 125 87 27 102 65 84 42.9 23.1 20972 (p) 24 56 31 12 14 13 39.6 23.4 20973 (p) 24 45 17 4 10 7 34.2 19.6 20975 (s) 52 83 45 44 24 34 45.3 20.3 20976 (s) 45 57 27 17 11 14 40.8 20.0 20977 (s) 33 35 9 5 4 4 46.1 18.5 20978 (s) 52 56 24 21 11 16 46.6 22.1 Continues….. 148 Table 89. Agronomic evaluation of a collection of Flemingia macrophylla in Quilichao. Preliminary data of two cuts (one in each season). Growth habit: =erect, s=semierect, p=prostrate. Treatment Height Diameter Regrowing Mean dry matter yields (g/pl) IVDMD Crude No. CIAT (cm) (cm) Points (No) Wet Dry Mean (%) Protein (%) 20979 (s) 48 76 38 27 25 26 38.9 21.2 20980 (s) 43 55 26 27 18 22 41.8 21.0 20982 (s) 49 61 28 26 23 25 41.0 19.9 21079 (s) 47 78 44 51 25 38 37.9 20.2 21080 (s) 41 58 13 32 11 21 39.2 15.5 21083 (e) 93 79 36 71 43 57 45.8 21.4 21086 (s) 27 29 4 NA 3 3 - - 21087 (s) 64 66 46 47 32 39 42.3 20.4 21090 (s) 88 106 48 135 66 100 50.0 21.3 21092 (s) 72 81 23 57 39 48 49.2 18.0 21241 (e) 133 93 27 134 66 100 36.2 20.2 21248 (e) 127 92 30 106 77 91 33.5 23.6 21249 (e) 129 104 34 167 85 126 40.9 22.0 21519 (e) 127 101 28 109 67 88 39.5 22.3 21529 (e) 132 102 31 145 71 108 42.0 23.1 21580 (e) 131 101 32 184 86 135 39.1 19.8 21982 (p) 19 62 38 26 11 19 42.1 20.9 21990 (p) 35 66 43 27 19 23 31.9 19.1 21991 (p) 29 52 24 13 10 11 37.5 22.6 21992 (p) 29 50 24 12 9 11 48.5 20.2 21993 (s) 42 77 45 34 24 29 43.9 19.9 21994 (p) 24 42 9 8 7 8 34.5 16.4 21995 (p) 29 50 26 11 8 9 40.8 19.6 21996 (p) 23 44 14 7 6 6 42.5 21.8 22058 (e) 84 58 13 41 29 35 37.2 18.5 22082 (s) 79 82 58 69 37 53 48.4 20.0 22087 (p) 27 51 17 15 4 10 40.3 17.8 22090 (s) 44 47 10 10 5 7 41.1 17.5 22285 (s) 43 75 42 32 21 27 38.7 20.4 22327 (s) 41 62 33 20 21 21 48.4 19.3 Mean 70 78 29 60 42 51 39.0 20.9 Range 15-146 21-122 4-58 0-184 1-91 1-135 31-51 16-24 Genetic analysis by molecular markers (AFLPs): Samples of 5 g of young leaves were taken of all C. argentea and F. macrophylla accessions and DNA was been extracted and quantified. To identify efficient primers for the AFLP analysis, 2 supposedly genetically contrasting accessions of each F. macrophylla and C. argentea (CIAT 21990 and 21529, and CIAT 18672 and 18516 respectively) were tested with different primer combinations and the resulting polymorphic bands were counted (Table 90). 149 Table 90. Polymorphic bands of different primer combinations for Flemingia macrophylla (accessions CIAT 21990 and 21529) and Cratylia argentea (accessions CIAT 18672 and 18516). Primer combination Polymorphic bands F. macrophylla C. argentea Total E-AAC / M-CAA n.a. n.a. n.a. E-AAG / M-CAA n.a. n.a. n.a. E-AAG / M-CAT 28 / 24 2 / 4 58 E-ACA / M-CAT 18 / 20 7 / 9 54 E-ACA / M-CTG 15 / 8 4 / 5 32 E-ACT / M-CTG 13 / 8 4 / 5 30 E-ACC / M-CAG 20 / 15 9 / 8 52 E-ACG / M-CAG 16 / 24 2 / 21 62 E-ACG / M-CAC 26 / 24 19 / 12 81 E-AGC / M-CTA 11 / 18 3 / 3 35 E-AGG / M-CTC 24 / 21 9 / 12 66 E-AAC / M-CTT 45 / 20 18 / 3 86 3.3.3 Agronomic characterization of a collection of Rhynchosia schomburgkii Contributors: M. Peters, P. Avila, L.H. Franco, B. Hincapié, and G. Ramírez (CIAT) Rationale From the evaluation of a range shrub legumes with tolerance to cool temperatures Rhynchosia schomburgkii emerged as one of the most promising species for higher altitude hillsides. Thus, we were interested in characterizing its potential feed value (Photo 16). Photo 16. Rhynchosia schomburgkii at Quilichao 150 Materials and Methods A total of 13 accessions of Rhynchosia schomburgkii, mostly originating from Colombia, were planted at Quilichao. Plants were transplanted into single-row plots, with 4 replications. Dry matter yield, drought tolerance and forage quality are the main parameters being measured. Results and discussion Results from last year had indicated that of the 13 accessions evaluated, CIAT 17918, 22134, 918 and 19235 showed the highest yields. This year we were interested in measuring quality parameters in the collection of R. schomburgkii as affected by seasonal variation. Results indicated that during the season with maximum rainfall, there were differences among accessions for IVDMD but not for CP (Table 91). In the drier period no significant (P>0.05) differences among accessions in terms of quality were recorded. However, season had a large effect on digestibility, but the effect was not the same for all accessions. Table 91. Fodder quality of accessions in a collection of Rhynchosia schomburgkii grown in Quilichao in Minimum and Maximum precipitation. Season Accession Minimum Maximum IVDMD CP IVDMD CP Tannins Soluble Bound 8582 42 22 42 20 4.81 0.78 19235 39 22 40 22 3.68 0.40 20800 38 21 38 21 2.49 0.84 17918 38 19 49 22 3.06 0.73 20456 38 19 52 23 3.31 0.68 22134 37 20 42 22 3.69 4.55 7389 36 22 44 23 2.95 0.67 7810 36 20 41 22 5.23 0.72 18490 36 21 40 22 3.44 0.92 918 31 19 38 22 5.85 0.72 LSD NS NS 5.7 NS (P <0.001) The concentration of condensed tannins measured in the wet season was relatively low (2.5 to 5.8%) and not as variable as IVDMD and CP. In general, our results show that, in the small collection of R. schomburgkii evaluated there is limited variability in CP and IVDMD which limits the scope for selecting genotypes based on quality. Activity 3.4 Selection of legumes for multipurpose use in different agroecosystems Highlights • Accessions of Vigna unguiculata with specific adaptation to acid or neutral soil and more broadly adapted accessions identified 151 • Research on Vigna unguiculata carried to Honduras and Nicaragua, Participatory evaluations in preparation • Lablab purpureus accessions with outstanding performance on neutral soils identified Progress towards achieving milestones • Suitability of Vigna unguiculata for acid and neutral soils defined Accessions were identified with specific and broad adaptation to variable soil pH and fertility conditions. • List of accessions of Vigna unguiculata for use as feed and/or green manure in Central America A core collection of Vigna unguiculata from IITA is now in Honduras and Nicaragua, for participatory evaluation. Seed multiplication of promising accessions is underway in Costa Rica. • Results on characterization of a core collection of Lablab purpureus in acid and neutral soils The Lablab purpureus accessions evaluated were well adapted to acid low fertility soils, eventhough productivity was much lower than on neutral higher fertility soils. However, there is intra-specific variation in adaptation to soil and climate conditions. Some accession have more specific adaptation while other accessions showed a more broad adaptation. The next step is to carry out more detailed studies with a limited number of accessions, focusing on small farmers in Colombia and Central America. For comparison, we are trying to obtain seed of available commercial cultivars (cv. Endurance Rongai, Highworth and Koala) from Australia. 3.4.1 Evaluation of a core collection of Vigna unguiculata for multipurpose uses in Colombia, Nicaragua and Honduras Contributors: M. Peters, Luis H. Franco, A. Schmidt, H. Cruz Flores, P. Avila, G. Ramírez, B. Hincapié, (CIAT), and B.B. Singh (IITA, Nigeria) CIAT projects: PE-2, PE-3 Quilichao and Palmira Rationale Cowpea (Vigna unguiculata) is utilized in the subhumid/semi-arid tropics of West Africa and India as a source of food and feed for livestock. Work of CIAT with a limited number of accessions had indicated potential of cowpea for soil improvement, but the utilization of cowpea in Latin America is so far limited. We visualize that, cowpea could be an alternative crop for the second planting season in the central hillsides region of Nicaragua and Honduras where the legume could provide not only higher grain yields as compared to common beans, but could also allow for a third crop in November/December in order to provide hay as animal feed in the dry season or contribute to soil fertility enhancement for the following maize crop. Adaptation to climatic and edaphic conditions, especially to water stress, are prerequisites for a successful development of a cowpea option for the traditional maize-bean cropping systems in Central America. It remains to be seen if cultural traditions allow for the inclusion of cowpeas in the daily menu of people in Central America. 152 A) Evaluation of cowpea in Quilichao and Palmira, Colombia Materials and Methods A core collection of 15 cowpea accessions was obtained from Dr. B.B. Singh, cowpea breeder of IITA and complemented with two local accessions from Colombia (cultivar Sinu) and Brazil (cultivar Verde Brasil). After initial experiments on acid soils (Annual report 2000), these accessions were again planted at CIAT’s Quilichao Research Station. Accessions were evaluated for grain and forage yield and their value as green manure for a succeeding maize crop (Photo 17). Photo 17. Vigna unguiculata in grass production phase at Quilichao Results and Discussion In the Quilichao site, all accessions established rapidly, reaching soil covers of >80%, 8 weeks after planting. At the time of incorporation into the soil (9 weeks after planting) all accessions were well established and vigorous. No significant differences (P>0.05) were found among accessions for DM yields (Table 92). However, significant differences (P<0.05) were found in maize dry matter production and in grain yield following the incorporation of cowpea accessions. Highest maize yields were recorded after green manuring with IT93K-573/5, with yields being 3.6 t/ha grain and almost 9 t of dry matter. In contrast, with no N grain and dry matter yields were 1.5 t and 4.1 t, respectively. Fertilizations higher than 80 kg N had a negative effect on maize grain and dry matter yields. Results confirm data obtained in the initial experiments (AR 2000). All green manure treatments except IT96D-759 led to higher maize yields than obtained with any level of nitrogen fertilizer applied. 153 Table 92. Dry matter yield (kg/ha) of cowpea green manure herbage and grain before soil incorporation and grain and dry matter yield of a following maize crop in Quilichao, 2nd phase. Cowpea Maize Accessions Herbage Grain DM Total (kg/ha) IT93K-573/5 3180 3619 8882 IT90K-284/2 2187 3558 8442 IT89KD-391 2387 3382 7576 IT95K-1088/4 2033 3350 7765 IT86D-716 2293 3290 8433 IT95K-1088/2 2213 3255 7779 IT86D-715 3313 3280 8308 IT6D-733 2867 3192 8993 IT96D-740 1940 3104 7321 IT90K-277/2 1913 3067 7520 IT93K-503/1 2047 2868 7219 IT86D-719 2393 2803 8238 IT93K-637/1 2613 2636 6338 IT89KD-288 3513 2558 6728 IT96D-759 1047 2331 5194 80N - 2405 5213 160N - 2104 4337 200N - 2094 4360 0N - 1487 4105 40N - 1478 3577 120 N - 1330 3785 LSD (P<0.05) NS 1862 3804 Forage quality of cowpea accessions in terms of CP, lignin, digestibility, P and Ca concentrations varied among accessions (Table 93). Nevertheless, with CP concentrations of 14-21 % and a digestibility of dry matter of 80% or more cowpea is also an excellent fodder for livestock (Table 93). Table 93. Fodder quality in accessions of Vigna unguiculata (cowpea) grown in Quilichao. Forage Grain Accessions Protein IVDMD Lignin P Ca N P K % IT86D-715 21 80 4.5 0.14 2.1 4.19 0.36 1.22 IT90K-277/2 19 82 2.5 0.12 2.1 3.00 0.28 1.00 IT93K-573/5 19 82 2.7 0.13 1.5 3.71 0.30 1.13 IT96D-740 18 83 5.9 0.13 1.6 3.20 0.33 1.20 IT90K-284/2 18 81 2.4 0.13 1.6 3.41 0.36 1.28 IT96D-733 17 84 4.4 0.12 1.5 3.37 0.33 1.22 IT86D-719 17 83 4.2 0.11 1.8 3.58 0.35 1.25 IT93K-673/1 17 85 2.7 0.13 1.5 3.47 0.34 1.25 IT93K-503/1 17 83 2.1 0.12 1.3 3.16 0.31 1.17 IT95K-1088/2 16 85 4.5 0.11 1.4 3.39 0.35 1.18 IT89KD-391 16 82 1.7 0.10 2.1 3.28 0.33 1.27 IT95K-1088/4 16 84 3.4 0.14 1.6 3.42 0.36 1.27 IT89KD-288 15 85 2.6 0.09 1.4 3.47 0.30 1.16 IT86D-716 14 86 5.6 0.10 1.3 3.78 0.35 1.3 LSD (P<0.05) 3.1 2.66 1.2 0.02 0.61 154 In a 3rd phase, cowpea accessions were sown in the same season in Quilichao and Palmira to compare the effect of climate and soil on performance and possibly identify accessions with broad adaptation, which is key for Central American Hillsisdes with highly variable soil and climatic conditions. The establishment of the accessions of cowpea included in the trial was slower in Palmira than in Quilichao, due to higher incidence of insects and weeds. In Quilichao, the incidence of pest and diseases was minimal, with the exception of a localized incidence of ants. Results showed that no differences in DM yields among accessions in the two sites (Table 94). However, mean dry matter yield in Quilichao (2229 kg/ha) was 30% higher than in Palmira (1752 kg/ha). In addition, we observed a G x E interaction in performance of accessions tested. For example, accessions IT86D-715 and IT89KD-391 had high DM yields on the acid soils in Quilichao, but were among the lowest yielding accessions on the more neutral fertile soils in Palmira. However, other accessions such a IT95K-1088/4 had high dry matter yields in the two sites. Table 94. Dry matter yield (kg/ha) of cowpea green manure herbage and grain before soil incorporation en Quilichao and Palmira, 2001. Accessions Quilichao Palmira DM Herbage (kg/ha) IT86D-715 3147 1280 IT89KD-288 2653 1873 IT6D-733 2567 1627 IT89KD-391 2413 1187 IT95K-1088/4 2373 2480 IT93K-503 2353 1627 IT96D-740 2230 1947 IT90K-277/2 2220 1307 IT86D-716 2187 1807 IT90K-284/2 2080 1900 IT93K-573/5 1993 1493 IT95K-1088/2 1827 2040 IT93K-637/1 1813 1927 IT86D-719 1773 2326 LSD (P<0.05) 1,130 1,252 B) Evaluation of cowpea in Nicaragua Materials and Methods A core collection of 19 accessions of Vigna unguiculata (Table 95) was established at the SOL SECO site (the Spanish acronym for Supermarket of technologies for hillsides – dry) in San Dionisio, Matagalpa, Nicaragua. The accessions were replicated three times in a randomized block design. Plots measured 5 x 2.5 m and seeds were sown at a distance of 0.25 m within a row and 0.5 m between rows. 155 Table 95. Accessions of Vigna unguiculata sown in San Dionisio/Nicaragua and Yorito/Honduras and Nicaragua as green manures for maize-based systems; Lablab purpureus DICTA was sown in Honduras only as a local check. Accessions Accessions Vigna unguiculata IT86D-277/2 Vigna unguiculata IT86D-715 Vigna unguiculata IT90K-284/2 Vigna unguiculata IT96D-740 Vigna unguiculata IT89KD-391 Vigna unguiculata IT95K-1088/4 Vigna unguiculata IT86D-716 Vigna unguiculata IT89KD-288 Vigna unguiculata IT93K-503/1 Vigna unguiculata IT93K-573/5 Vigna unguiculata IT93K-637/1 Vigna unguiculata CIDDICO1 Vigna unguiculata IT86D-719 Vigna unguiculata CIDDICO2 Vigna unguiculata IT95K-1088/2 Vigna unguiculata CIDDICO3 Vigna unguiculata IT6D-733 Lablab purpureus DICTA Evaluations will include seed emergence, ground cover, plant height, plant vigour, biomass/grain production flowering patterns, and incidence of pest and disease. Local farmers will be invited to participate in the evaluation of the core collections and soil fertility enhancement effects will be measured through the planting of a maize crop at the onset of the next wet season and comparing maize yields with N-fertilized plots. Expected results The selection of superior accessions based on agronomic performance on a farmer criteria will provide a clear indication on the potential of Vigna unguiculata in farming systems found in Hillsides of Central America. Further evaluations with the selected accessions will be necessary in order to optimize management techniques. C) Evaluation of cowpea in Honduras In 2001, an experiment was established in the SOL Yorito to evaluate a core collection of cowpea (Table 96). In this case the experiment was complemented with the addition of Lablab purpureus DICTA as a control. In Yorito, the focus is again on selecting cowpeas for green manures in maize-based systems for soils with neutral to alkaline pH. Significant (P<0.0001) differences among accessions were found for DM yield. The highest biomass production was recorded with CIDICCO3 (6.2 t of DM/ha) and IT90K-284/2 (6.1 t of DM/ha). In general, the ranking of accessions compares favourably with results obtained on neutral soils in Palmira though yields are much higher in Honduras. 156 Table 96. Dry matter yields of Vigna unguiculata (cowpea) genotypes before soil incorporation before a maize crop in Yorito, Honduras. DM yield Soil cover (%) kg/ha CIDICCO3 93 6212 IT90K-284/2 90 6123 CIDICCO1 83 5282 Lablab purpureus DICTA 85 5230 IT96D-740 72 5112 IT93K-637/1 72 5101 IT86D-716 67 4944 IT95K-1088/2 72 4042 IT95K-1088/4 68 3923 IT6D-733 50 3827 IT93K-503/1 60 3672 CIDICCO2 78 3521 IT93K-573/5 50 2926 IT89KD-391 43 2867 IT89KD-288 53 2734 IT86D-719 38 2381 IT86D-715 37 2175 IT90K-277/2 33 1754 LSD (P<0.05) 2014 3.4.2 Evaluation of core collection of Lablab purpureus for multipurpose uses (Quilichao and Palmira) Contributors: M. Peters, L. H. Franco, B. Hincapié, and G. Ramírez (CIAT) Rationale Lablab purpureus is a free seeding, fast growing or short-term perennial legume, with widespread use through the tropics as a fodder plant. In Africa the use of Lablab for human consumption is also common. The origin of the Lablab germplasm currently utilized is mainly Eastern/Southern Africa and Asia. In addition, it is well documented that Lablab purpureus is best adapted to lower altitudes and to areas with rainfall regimes of 750–2000 mm/year. This species grows in a variety of soils, but the ideal pH for growing Lablab is reported to be between 5.0 and 7.5. In order to evaluate the potential of Lablab in tropical America, we obtained a collection available at ILRI/CSIRO. Our main objective with the collection is to select accessions with broad adaptation to different soils and climate conditions in tropical America. However, of immediate interest is to evaluate the Lablab collection in acid and neutral soils to define niches of Lablab for green manure and fodder (especially for hay and silage or deferred feed), with emphasis on Central America where soils are highly variable in pH. Materials and Methods A total of 44 accessions of Lablab purpureus were initially sown on an acid soil (pH 4.0) in the Quilichao Research Station for seed multiplication. In 2001, 42 and 25 accessions were planted for agronomic evaluation in neutral (Palmira), and acid soil (Quilichao), respectively (Photo 18). 157 Photo 18. Seed of Lablab purpureus at Quilichao Results and Discussion Results of the agronomic evaluations are shown in Tables 97 and 98. In Quilichao (Table 97), accessions 14442, I 14411, I 14437, 76996, 21603 and 99985 were the fastest to establish, with soil cover of >95 % and vigour ratings of 4 to 5, 12 weeks after sowing. Of the 25 accessions sown 15 (60%) had a soil cover above the mean of the experiment (87% soil cover). As expected, early flowering accessions were less productive than late flowering accessions. The highest yields were recorded with accessions CIAT 34777, 52535 and 21603, and the lowest yields were recorded for T 52508, 17192, I-6536 and I-11613. However, it is interesting to note that accession I- 6536, which had very low yields at 12 weeks, became one of the most productive accessions, 4 weeks later. Plant vigour and yields of Lablab in Palmira (Table 98) were higher than in Quilichao. In this site significant (P<0.05) differences among accessions were found for DM yield and soil cover. The accessions with fastest soil cover and ability to compete with weeds measured 8 weeks after planting were 14442, I-11630, I-14437, 29398, 76996, 34777, I-6533, I-14411, L-987 and 106494. In general, the accessions with the best adaptation across different soil and climate conditions were 34777, 96924, 21603, I-11630, I-14411, I-14441, 67639 and 52535. 158 Table 97. Dry matter yield (kg/ha) and soil cover (%) of Lablab purpureus herbage in Quilichao, 2001. Accessions Vigour Cover (%) DM ( kg/Ha ) 1 a 5 12 Weeks 16 Weeks 12 Weeks 16 Weeks 34777 4 83 50 2447 2153 52535 4 77 57 2440 1973 21603 5 97 90 2327 2367 I-11632 3 73 50 2067 1453 76998 4 93 77 2060 1607 99985 5 95 97 2027 1927 106494 5 87 93 2013 1153 36903 3 73 53 1920 1453 I-14411 5 100 93 1913 1573 100602 4 80 87 1887 1627 I-14437 4 98 83 1853 2447 I-11630 5 70 100 1840 2093 14442 5 100 97 1840 2667 67639 4 93 80 1820 2313 CQ-2975 4 93 87 1807 2046 106548 4 90 93 1793 1833 81626 3 80 60 1707 2073 106500 4 93 90 1640 1560 76996 5 98 90 1560 1873 I-14441 4 90 73 1493 2067 I-6533 3 77 70 1426 1273 I-11613 4 92 93 1280 1387 I-6536 5 93 97 1213 1980 17192 2 80 77 1160 1060 52508 3 67 50 1147 1053 22183 . 1453 LSD 22.5 16.4 878 NS (P<0.05) 159 Table 98. Dry matter yield (kg/ha) and soil cover (%) of Lablab purpureus herbage in Palmira, 2001. Treatment Vigour cover(%) DM kg/ha 1 a 5 8 weeks 13 Weeks 8 Weeks 13 Weeks 35894 2 93 70 3493 9067 I-14437 4 97 97 3280 7760 I-11615 2 87 60 3293 7340 34777 4 97 77 3380 6933 96924 2 83 77 2840 6927 21603 4 93 100 3760 6847 I-11630 5 100 100 4807 6740 2160 3 87 77 2707 6707 I-14411 4 97 93 3080 6587 I-14441 4 87 95 2420 6547 67639 4 87 97 2780 6533 29398 4 97 93 3447 6527 L-987 5 77 100 2367 6407 52535 3 93 73 3360 6340 I-6533 4 97 87 3193 6300 52544 3 87 83 2413 6187 106494 4 97 100 2440 6120 100602 3 87 90 2113 5660 L-1683 5 77 100 2367 5500 76998 4 93 93 3433 5480 76996 4 97 93 3213 5440 81626 2 83 53 2487 5433 CQ-2975 4 80 90 2387 5407 17197 3 57 80 1560 5193 I-6536 3 73 90 2000 5173 I-11613 4 87 90 2780 5027 I-11632 2 83 63 2793 5007 36903 3 93 60 2660 4900 106548 4 77 90 2067 4747 14442 5 100 97 2367 4640 99985 3 90 63 3113 4587 I-6930 2 83 57 2247 4527 52508 2 77 47 2473 3927 22183 4 83 93 2100 3853 106500 3 73 100 1373 3827 69498 2 63 73 1387 3747 17196 2 50 77 867 2987 51564 2 40 73 500 2880 17193 1 43 60 680 2020 17192 1 23 43 393 1147 17195 1 17 73 187 1140 17189 1 20 20 247 853 LSD (P<0.05) 35.2 47.3 2514 5557 160