Desmodium heterocarpon (L.) DC.
subsp. ovalifolium (Prain) Ohashi

Leguminosae

Synonyms

• Desmodium polycarpum DC. var. ovalifolium Wallich ex Prain (1897)
• Desmodium ovalifolium Wallich ex Merrill (1910)
• Desmodium heterocarpon (L.) DC. ssp. heterocarpon var. ovalifolium (Wallich ex Prain) Rugayah (1987)

[The profile is based on: Schmidt, A. & Schultze-Kraft, R. 2001. Desmodium ovalifolium – a review. Tropical Grasslands: submitted].

Common names

Desmodium ovalifolium, desmodium, desmodio, khonthi din (Thailand), tràng qua xoan (Vietnam)

Description

The following species description is taken from Schultze-Kraft (1992): A creeping, stoloniferous herb, in dense stands or under competition ascending up to 1 m. Stems many-branched, glabrous except for silky pubescence on young (apical) portions; old stems woody at the base. Leaves 1- and 3-foliolate, in young plants always 1-foliolate; leaflets variable, mostly ovate or broad-elliptical, sometimes round or obovate; terminal leaflet larger than laterals; 1-foliolate leaves and terminal leaflets on adult plants on average 3-4.5 cm x 1.5-3 cm; leaflets coriaceous and without any markings, glabrous and glossy on upper surface, whitish pubescent on lower surface, sometimes only along the central vein. Inflorescence a densely flowered raceme, 2-5 cm x 1-2 cm; flower papilionaceous, small, with an obovoid standard about 6 mm x 4 mm, purple to dark-pink during anthesis, bluish when wilting. Pod erect or falcate, mostly densely pubescent, dehiscent, comprising 2-8 almost quadrate articles 2.5-3.5 mm long. Seeds yellow when ripe.

Taxonomy

"Desmodium ovalifolium" is the abbreviated name (used in non-botanical literature) for a distinct, taxonomic Desmodium entity which earlier had been recognized at the species level (Merrill, 1910: D. ovalifolium Wall.), then rejected as an independent taxon and considered rather as part of D. heterocarpon (L.) DC. (Ohashi, 1973), subsequently recognized as a botanical variety (Rugayah, 1987: D. heterocarpon subsp. heterocarpon var. ovalifolium (Wall. ex Prain) Rugayah), and finally recognized at the sub-species level. The taxonomically valid name is at present Desmodium heterocarpon (L.) DC. subsp. ovalifolium (Prain) Ohashi (Ohashi, 1991). Desmodium ovalifolium belongs to the tribe Desmodieae of the Fabaceae subfamily Papilionoideae.

Because of plant morphological characteristics, D. ovalifolium is easily distinguishable from the typical D. heterocarpon subsp. heterocarpon. However, overlapping between the two taxa occurs, and there are indications of some degree of cross-pollination in this mainly self-pollinated species (Schultze-Kraft & Benavides, 1988; Ohashi, 1991). This is supported by successful hybridisation within the D. heterocarpon - D. ovalifolium species complex whose chromosome number is 2n = 22 (Quesenberry et al., 1989).

Geographic distribution and genetic resources

The geographic distribution of D. ovalifolium extends over moist-subhumid to humid (1200-4500 mm rainfall a^-1) Southeast Asia: Thailand, Cambodia, Laos, Vietnam, Malaysia, Indonesia (Sumatra, East Kalimantan, Sulawesi, Java) and the Philippines (Ohashi, 1991). The species occurs generally at low elevations. Noteworthy exceptions occur between latitudes 20°30' N and 04° S in environments that are frequently characterized by some degree of shade. As a result of a number of collecting missions under the auspices of the Centro Internacional de Agricultura Tropical (CIAT), Cali, Colombia, between 1979 and 1992 a total of 123 germplasm accessions were collected in Indonesia, Malaysia, Thailand and Vietnam (Schultze-Kraft et al., 1989, 1993). At present, a collection of 160 accessions is maintained in the CIAT genebank; basic passport data of accessions can be accessed on the Internet (http://singer.cgiar.org/).

The collected germplasm has shown little morphological variation in small-plot field nurseries but considerable diversity regarding establishment vigour, flowering time, dry matter and seed production, drought tolerance, and forage quality (Schultze-Kraft & Benavides, 1988). Prospects for further describing the genetic diversity (and eventually identify accession duplicates) by means of polyacrylamide gel electrophoresis (PAGE) were initially promising (Hussain et al., 1987). Several suitable isozymes were identified but at the genotype level they were subsequently not sufficiently discriminatory (Klein, 1997).

Adaptation

D. ovalifolium is adapted to acid, low-fertility soils and has low nutrient requirements (e.g. Sánchez & Salinas, 1981; Suárez et al., 1985; Urdaneta et al., 1985; Ahmad, 1986; Silva del A., 1986; Costa et al., 1988). The species showed also good adaptation to water-saturated and periodically flooded savannas (CIAT, 1986a, 1989a) and was generally considered suitable for high rainfall regions (moist-subhumid to humid tropics) with some drought tolerance (Dutra et al., 1980; Spain, 1981; CIAT, 1987a; Miles & Lapointe, 1992; Rao et al., 1992). For more detailed and location specific information on the adaptation potential of D. ovalifolium (mainly accession CIAT 350) in different agro-ecological zones, the reader is referred to the publications of Franco et al. (1990, 1992a, 1992b).

Through its aggressive, stoloniferous growth habit, D. ovalifolium is one of the few legumes compatible with mat-forming grasses like Brachiaria decumbens (Grof, 1982) and was reported to depress weeds effectively in pastures and plantations (CIAT, 1989b, 2000). In contrast to numerous reports mentioning good persistence of the legume (e.g. CIAT, 1985a; Villarreal & Chávez, 1991; Ortega & Samudio, 1997), even under severe animal trampling conditions (García, 1990), D. ovalifolium seems to be affected by soil compaction due to high stocking rates under high rainfall conditions and short regrowth periods (Maldonado & Velásquez, 1990; Reátegui et al., 1990). Furthermore, it was found sensitive to burning (Grof, 1980). To what extent its shallow root system (Ayarza, 1988) has negative impact on persistence is not clear, but Bishop (1983) regards this as an advantage for using the species in silvopastoral systems. Experiments in West Africa suggest the unsuitability of the legume for most of this region (Hairiah & Noordwijk, 1989; RABAOC, 1995; Smith, 1995). This lack of adaptation to drier environments was confirmed by a study using information of origin analysed with FLORAMAP (Jones & Gladkov, 1999). The study identified potential collection sites in South-East Asia and sites for long-term survival (Jones et al., 2000).

Establishment

Ferguson (1992) regards D. ovalifolium as easy to establish and highly persistent. Spot-sowing without major soil preparation is possible (CIAT, 1984a); this is particularly important for fragile soils (Suárez & Cardona, 1993). An additional way of seed distribution can be through animal faeces when seeds are mixed with mineral salt supplementation. Furthermore, D. ovalifolium was found suitable for establishment via mixed cropping with upland rice in the Llanos Orientales (Eastern lowlands) of Colombia (Schultze-Kraft & Cárdenas, 1993). In case of seed shortage, it can be propagated vegetatively; adequate techniques are described by Grof et al. (1981). Surface-sowing of D. ovalifolium macro-pellets into burnt savanna is another possible method of establishment (Ogawa et al., 1990; Nada et al., 1992). Recommended seeding rates range from 0.5 kg ha^-1 for pasture associations to 5 kg ha^-1 for cover crop purposes (Spain, 1982; Bishop, 1983; Aguila, 1992); trials to define optimal seeding rates in rubber and oil palm plantations are underway, initial results indicating seeding rates of 1-2 kg ha^-1 as suitable (M. Peters, unpublished). A general disadvantage of the legume is the repeatedly reported slow germination and establishment (e.g. Ridzwan & Sariam, 1982; López & Silva del A., 1986; Domínguez-Valenzuela, 1990). A summary of various establishment aspects of pasture legumes, including D. ovalifolium, can be found in Lascano & Spain (1991).

Agronomic performance

In regard to dry matter production, Grof (1984) recorded at Carimagua, Llanos Orientales of Colombia, legume yields of 7.3, 7.2 and 3.98 t ha^-1 a^-1 in associations of D. ovalifolium with Brachiaria humidicola, B. decumbens and Andropogon gayanus, respectively. In RIEPT network evaluation experiments, legume dry matter production (12 week regrowth) of accession CIAT 350 varied from 0.5-6.5 t ha^-1. The respective location specific results are compiled in the proceedings of various RIEPT meetings (Pizarro, 1983, 1985, 1988, 1992; Keller-Grein, 1990). More aggregated information is also available in Franco et al. (1990, 1992a, 1992b). In general, D. ovalifolium can be considered as a productive legume in high-rainfall areas; it is highly compatible in associations with important tropical grasses.

Fertilization

The species responds readily to fertilization. Application of lime and sulphur increased dry matter yields (Pinzón et al., 1980; Yaringaño, 1983). Amounts of 500 kg ha^-1 lime and 20 kg ha^-1 sulphur are recommended for pasture establishment (Sánchez & Salinas, 1981). A dramatic response to magnesium was recorded by Spain (1982). Boron and phosphorus application also had a positive effect on dry matter production (Salinas et al., 1987). Phosphorus requirements of D. ovalifolium are classified as intermediate (Costa, 1988). For more detailed reading, Pérez (1997) provides a summary of fertilizer effects on yields of D. ovalifolium on different soils in the Llanos of Colombia.

Nutritive value

In contrast to the remarkable agronomic performance of D. ovalifolium in high rainfall areas, the nutritive value of the species falls behind other tropical pasture legumes such as Centrosema spp. (Lascano et al., 1990) or Arachis pintoi (Lascano, 1994). This is mainly attributed to the high condensed tannin (CT) content of the legume with subsequently reduced in vitro digestibility (IVDMD), low palatability and animal intake (Carulla, 1994; Lascano et al., 1995). Data on CT contents of D. ovalifolium, however, are limited. Abaunza et al. (1991) reported a CT content of 17% in D. ovalifolium accession CIAT 350 at Quilichao, Colombia. This value is lower than those observed in former collection screenings at the same location where CT contents ranged from 19-43% (CIAT, 1984b; Schultze-Kraft & Benavides, 1988). Leaf tissue analysis from Carimagua revealed CT contents which varied within and between seasons (16-25%) (CIAT, 1984b). These data were obtained through the vanillin-HCl method (catechin standard) (Price et al., 1978) and are difficult to compare with more recent studies using different tannin analysis methods. Carulla (1994) reported CT contents of 4-7% using the butanol-HCl method and D. ovalifolium tannins as standards (Asquith & Butler, 1985; Terrill et al., 1992). Furthermore, it is unclear if all results were obtained with adequate sample preservation methods. Valerio (1990) and Cano et al. (1994) found differences in tannin contents in relationship to sample preservation and recommended freeze-drying as appropriate for legumes containing tannins.

Published information on IVDMD of D. ovalifolium (CIAT 350) shows also a large variability across locations (Table 1). Values range from 29-56%, which is below the mean value for tropical legumes (56.6%) reported by Minson & Wilson (1980). Schultze-Kraft & Benavides (1988) likewise found high variability in nutritive value among accessions of D. ovalifolium under the same environmental conditions. Crude protein contents ranged from 12-20%, phosphorus and calcium concentrations from 0.10-0.16% and 0.27-0.57%, respectively. Although these values are low compared to other tropical legumes (Ara, 1987), they are well in the range of results reported in literature for D. ovalifolium accession CIAT 350 from different locations (e.g. Giraldo et al., 1989; Abaunza et al., 1991). Laredo (1985) reported fibre contents (acid and neutral detergent fibre) in the range of 40-50% for material from the Llanos of Colombia.

Bdic = Brachiaria dictyoneura; Bh = Brachiaria humidicola; Ap = Arachis pintoi; Pp = Pueraria phaseoloides

The variability in nutritive value and especially in CT content among genotypes, locations and seasons indicate the existence of pronounced genotype x environment interactions in D. ovalifolium determining forage quality. This is further supported by the work of Salinas & Lascano (1983) who observed reduced CT contents in D. ovalifolium under increased fertilizer application, especially sulphur, at Carimagua.

Animal production

Animal production data provide evidence for an even more complex interaction situation including the factors associated grass species and pasture management. Table 2 shows a large variability in cattle liveweight gains (LWG) at different locations in South America under a range of stocking rates and associated grasses. Mixtures with Brachiaria humidicola seem to be particularly productive. The contrast between very high and very low LWG is noteworthy. This is corroborated with additional reports on both good performance of D. ovalifolium/grass associations and very low or even negative LWG: While, for example, Lascano et al. (1982) report a 400% daily LWG increase in a B. humidicola-D. ovalifolium association compared to the pure grass pasture (400-100 g an^-1 d^-1, respectively), Gareca (1986) measured LWG of the same association on farms in the Llanos Orientales of Colombia which were even lower than those of native savanna. Data obtained earlier at CIAT (1984c) indicate cattle LWG of -21 g an^-1 d^-1 in an association with Andropogon gayanus at Carimagua.

Milk production is regarded as more sensitive to forage quality. A review of the respective data reveals a similar contrasting picture to that for cattle LWG. Sanabria & Pabón (1999) measured a 10% increase in milk production when cows grazed a B. decumbens-D. ovalifolium mixture as compared with pure grass in the Amazon region of Colombia. The data of Heurck (1990), however, indicate slightly lower milk production on D. ovalifolium in association with Cynodon nlemfuensis compared to the pure grass pasture in Costa Rica. Compared to other legumes (e.g. Stylosanthes guianensis, Centrosema macrocarpum or Arachis pintoi), there is evidence that the milk production potential of D. ovalifolium-grass associations is inferior (Heurck, 1990; González Chau, 1992; Caruzo & Vela, 1995).

Pasture management

Another controversial aspect of D. ovalifolium/grass pastures, which has been related to its nutritive value, is the tendency of the legume to dominate pastures over time. Toro (1990) reports legume dominance in pastures with D. ovalifolium accession CIAT 13089 at Carimagua. Muñoz & Costales (1985) made similar observations with CIAT 350 in Napo, Ecuador. Additional examples for legume dominance are given by CATIE (1989b) and Campero (1994). Several authors, however, relate this tendency to inadequate pasture management, especially to low stocking rates (e.g. Tergas et al., 1981; CIAT, 1985c; Vela & Flores Mere, 1996). Vela (1994) recommends high stocking rates (4 animals ha^-1) for pastures containing D. ovalifolium at Pucallpa, Peru, in order to maintain the grass-legume balance. Martins da Silva (1982) sustained an adequate grass-D. ovalifolium balance through stocking rates of 3.5 animals ha^-1 at Carimagua. It was assumed by Bishop et al. (1992) that pasture management has to keep D. ovalifolium plants in young regrowth stages, since older tissue is not grazed. Maeno (1985) and Ferrufino (1989) associated increased consumption of D. ovalifolium with plant renewal at Quilichao, Colombia and Chiripiri, Bolivia, respectively. Higher IVDMD and increased legume intake are reported with high stocking rates by CIAT (1982a, 1986b). High stocking rates are also seen as a way to eradicate the legume from pastures (Cadisch et al., 1996). Fire might be an additional possibility to control the legume, since D. ovalifolium is susceptible to fire (Reátegui et al., 1995). Unfortunately, only few data on forage quality related to grazing and pasture dominance are available. Such data, however, are needed to review the contrasting observations concerning the relationship between high tannin contents and legume dominance in pastures.

Diseases

The use of D. ovalifolium may be limited by several diseases and pests. Pink disease (Phanerochaete salmonicolor) is currently restricted to Malaysia where it is serious only on D. ovalifolium in rubber plantations. Mostly young branches of the plants are affected and show salmon-pink mycelial encrustation before dying (Lenné & Stanton, 1990). Wart disease or false rust, caused by the fungus Synchytrium desmodii, was first recorded in Kalutara district in Sri Lanka in 1952 in rubber plantations (Munasinghe, 1955) and was introduced to South America through the importation of D. ovalifolium seed (Lenné, 1985). The disease was recognised for the first time in South America in 1981 in evaluation plots at Carimagua (CIAT, 1982b). Although adult plants generally regrow healthy tissue even after severe attacks of false rust, the fungus decreases herbage and seed production, and may also dramatically reduce seedling survival and thus long-term persistence of the legume. Seed treatment with concentrated hydrochloric acid in order to control further distribution of the disease by seed failed to kill resting spores of the fungus. Control through application of several fungicides was not successful either (Lenné et al., 1990). A geographically widespread disease, which also affects D. ovalifolium, is little leaf, caused by polymorphic, mycoplasma-like organisms (MLOs). However, the disease, also known as phyllody or witches’ broom, is generally only recorded under experimental conditions and is therefore not a major limitation in production systems (CIAT, 1984e). For an extended and detailed review of diseases of the genus Desmodium, the reader is referred to Lenné (1994).

Pests

Due to their widespread distribution, root-knot nematodes are considered by plant pathologists as a serious threat to D. ovalifolium in Central and South America (Trutmann, 1994). Infected plants become stunted, chlorotic and wilted, some defoliate and die (Lenné, 1981a). Meloidogyne javanica, M. incognita, M. arenaria and M. hapla are reported to affect D. ovalifolium (Lenné & Stanton, 1990), but there seems to be large variation among locations, years, accessions and the subsequent effect on plant persistence. Lenné et al. (1990) considered the role of nematodes as important only for the well-drained savanna ecosystem (such as at Carimagua, Llanos of Colombia). Most reports are based on experiences from germplasm screening sites like Carimagua and Pucallpa, Peru (e.g. CIAT, 1984f, 1988b). There seems to be evidence for a build-up of nematode populations at these sites showing increased plant infection over the years. In 1982, accession CIAT 350 was reported as only slightly affected at Carimagua (CIAT, 1984e) and still characterised as of good resistance in 1983 (CIAT, 1984g). Two years later the same accession was described as severely affected by nematodes at this major germplasm screening site (CIAT, 1986a).

Large variation was found by screening D. ovalifolium accessions for their tolerance/resistance not only to Meloidogyne spp., but also to Pterotylenchus cecidogenus, a stem gall nematode (CIAT, 1984g, 1986c, 1989a; Hernández & López, 1985; Sasser et al., 1987; Stanton et al., 1990). This above-ground nematode produces galls on stems, causing disruption of the vascular systems and eventual death of D. ovalifolium plants (Siddiqi & Lenné, 1984). It has only been recorded in the Llanos of Colombia and near Porto Velho, Brazil (Stanton & Torres, 1989). Stanton (1994) claims that pure stands of D. ovalifolium may be killed by the nematode within two years, while subsequent survival of seedlings and shoot and root growth are also reduced. From several accessions classified as nematode resistant, the high yielding accession CIAT 13089 was selected because of its additional tolerance to Synchytrium desmodii and it subsequently replaced CIAT 350 in further on-station and on-farm experiments (CIAT, 1990a). No reports are available on factors determining this resistance. Quesenberry & Moon (1995) used selected D. heterocarpon materials to enhance root-knot nematode tolerance. For more extensive information on nematodes in tropical legumes the reader is referred to Stanton (1994).

The potential of chemical control such as pregermination treatment of D. ovalifolium seed or application on seedlings was only studied in pot trials and showed mainly short term effects (Stanton & Torres, 1989). Remarkable are reports from agronomists on the good recuperation of D. ovalifolium screening plots from soil seed banks after severe nematode and false rust attacks at Carimagua (CIAT, 1985a, 1985d). When grown in mixtures with grasses only slight damage of M. javanica on D. ovalifolium CIAT 350 was observed (Lenné, 1981b), which, from the agronomic point of view, was considered insignificant, questioning the replacement of this accession by CIAT 13089 (CIAT, 1984a). Nevertheless, since D. ovalifolium has to be considered as host for nematodes, its use as cover crop in banana plantations could be limited as nematode populations (e.g. Radopholus similis) may be maintained (Ternisien, 1989). Also the larvae of the beetle Eutheola sp., commonly known as chiza in the Cauca area in South Colombia, may be seen as a limitation (CIAT, 1989c; Lascano et al., 1991). Although leaf sucking and eating insects were observed in D. ovalifolium, they do not constitute a real threat for the species. A summary on insect pests associated with forages in tropical America was given by Calderón & Arango (1985). Except for the controversial issue of nematode importance and the fungus Synchytrium desmodii, the legume can be considered as relatively free of diseases and pests (CIAT, 1984e; Béreau, 1986).

Seed production

Since all accessions of D. ovalifolium represent undomesticated populations, they retain many wild-plant characteristics, which aid their natural spread but pose difficulties for commercial seed production (Loch & Ferguson, 1999). Very little detailed knowledge is available about their flowering behaviour and the identification of regions suitable for seed production. Table 3 gives an overview of seed yields obtained at different locations. Seed production varied with location, year and genotype, as did the time from flowering to seed harvest. Yields ranged from 0.1-440 kg ha^-1 a^-1. Favourable seed production conditions were found in regions of higher latitude with high moisture availability throughout the year, e.g. in Chimore, Bolivia, 18° S, 226 masl, 21-26 °C, no frost risk, 4000 mm rainfall with 1000 mm in the period of minimum precipitation (Ferguson et al., 1983). The legume is very sensitive to moisture stress which may lead to low, erratic or no flowering at all (Andrade et al., 1983; Thomas, 1986). Flowering may be extended over the whole year, if adequate moisture is available, favouring the legume's self-generation under various climatic conditions, but undesirable for commercial seed production (Sobrinho, 1982).

Thousand-seed weights of D. ovalifolium range from 1.7-2.2 g (~500,000 seeds kg^-1) (Rotar & Urata, 1966). When dried to low moisture contents of 4-8%, seeds can be stored for long periods. Little loss in germination was observed after storage for 16 years at temperatures as low as -12 °C (Bass, 1984). Similar to other tropical legumes the hardseededness of D. ovalifolium seed has to be broken prior to sowing. High temperature treatment with hot water and scarification with sulphuric acid are recommended (Rojas & Herrera, 1989). Crop management guidelines for seed production are not documented, apart from information on chemical weed control (Ferguson & Sánchez, 1984).

Commercial cultivars

Desmodium ovalifolium seed is commercially available, mainly in the form of an unnamed variety traded in Southeast Asia and known to researchers as accession CIAT 350. From this material, a commercial cultivar (cv. Itabela) was developed and officially released in Brazil for use as a forage legume in the humid tropics of the state Bahia (CEPLAC, 1990). In Colombia, the release of the same genotype as cultivar "Macagual" is currently being prepared (Velásquez et al., in preparation).

Use as cover crop

The potential of D. ovalifolium for integration into tropical production systems is not only based on the already described agronomic characteristics concerning its use as a pasture plant. The legume's attributes favour also its utilization as a cover crop and thus as a multi-purpose legume. The first publications on D. ovalifolium dealt exclusively with its role as soil cover under rubber in Sri Lanka (Changeasfrera, 1954; Munasinghe, 1955) and Malaysia (Anon. 1960; John, 1963). Edgar (1958) reports D. ovalifolium as part of the traditional cover plant mixtures in Southeast Asia. High persistence through outstanding adaptation to shade environments and its non-climbing growth habit explain its particular suitability for plantation agriculture (Wong et al., 1985). Wahab & Ahmad (1984) observed even better dry matter production under shade than in full sunlight. D. ovalifolium was found the most shade tolerant species producing dry matter of 5 t ha^-1 a^-1 in the experiments of Chen & Sariam (1984). Likewise, in South and Central America the legume is increasingly used in plantation agriculture because of its good ground cover potential (Béreau, 1986; CIAT, 1987b) and dry matter production under shade (Achío, 1997). In the southern part of Zulia state, Venezuela, D. ovalifolium is used as both cover crop and pasture legume on more than 10,000 ha (I. Urdaneta, CORPOZULIA, 1998, pers. communication). In this context, seed production of D. ovalifolium in plantations is attracting growing interest among farmers.

Extensive weed suppression through D. ovalifolium compensates for the species' slow establishment from seeds (Ferrufino & Ovando, 1988; CIAT, 2000). Due to this latter restriction, D. ovalifolium is used in legume mixtures with faster establishing species like Centrosema pubescens, Calopogonium mucunoides, C. caeruleum, Mimosa invisa and Pueraria phaseoloides in young plantations. As light penetration through the tree canopy will decrease over time, only D. ovalifolium will persist because of its shade tolerance (Watson, 1963; Jayasinghe, 1991). Chee (1982) reports persistence of more than 10 years under rubber. A possibility to shorten the establishment period is vegetative propagation (Herrera, 1996).

Desmodium ovalifolium is not only recommended for oil palm or rubber, but also for a number of other perennial plantations such as coffee, tea, banana or macadamia nuts (e.g. Suárez, 1992; Sánchez, 1993, Humphreys, 1994; Bradshaw & Lanini, 1995). However, the evaluation of the resulting cover plant/main crop interactions is controversial. Pérez et al. (1993) observed negative effects on growth and fruit yields of Bactris gasipaes (peach palm) at Yurimaguas, Peru, because of alleged competition for nitrogen. Although D. ovalifolium was found to suppress weeds effectively (CIAT, 1990c) and serve as a trap crop for nematodes in coffee orchards (Vallejos, 1993), there are neither clear studies on further possible interactions nor large data sets on the economic viability of the technology under different ecological and market conditions (Bradshaw et al., 1992). A positive example of the use of D. ovalifolium in coffee is the adoption of a slash-and-mulch management of accession CIAT 350 to reduce weeding costs (reduction from 4 to 2 cleanings a^-1) and to improve soil fertility with robusta coffee production in the Coca region of Ecuador (Peck & Bishop, 1992): Every 6 months Desmodium is cut back to ground level and the mulch material left in place to decompose. This forms a layer of organic material favourable for mycorrhizae development, on which coffee growth is strongly dependent, and improves nutrient cycling. The management technique is economically interesting (CIAT, 1991b), and would have been even more adopted by farmers if more palatable genotypes for livestock integration (goats) had been available for the region (Lascano & Pezo, 1994). Nevertheless, Manidool (1985) and Domínguez-Valenzuela (1990) favour the use of D. ovalifolium in agroforestry/silvopastoral systems in the humid tropics, either in livestock systems including plantation grazing of grass-legume mixtures, or as a cover crop in plantation agriculture. The use of D. ovalifolium in improved fallow systems for crop production is suggested not to be advisable due to the tendency of the legume to weediness in later crop cycles through the build-up of soil seed banks (Staver, 1989). Recently, D. ovalifolium was used successfully for the rehabilitation of erosion galleys in South Colombia (Villada et al., 1997).

Nitrogen fixation

Although the species is not considered a particularly efficient N₂ fixer (Cadisch et al., 1989; Viera-Vargas et al., 1995), it can contribute >200 kg N ha^-1 a^-1 (Boddey et al., 1995; Valles et al., 1996), if effectively nodulated by the required specific Bradyrhizobium strain (Giller & Wilson, 1991). As a result of extensive evaluations, D. ovalifolium CIAT strain 4099 was recommended by CIAT (Franco et al., 1993). Inoculation efficiency can be increased by molybdenum fertilization (Balaguera, 1986) and inoculum application as water solution on young seedlings (CIAT, 1988c). A decrease in BNF of D. ovalifolium has been observed under increasing stocking rates (Boddey et al., 1995) and under shade conditions (Reynolds, 1995), but BNF seems to be enhanced through P and K applications (Cadisch et al., 1989).

Nutrient cycling

Nitrogen fixed by D. ovalifolium may contribute to the N economy and productivity of agricultural systems mainly in two ways: directly through the consumption of the legume tissue by animals, and indirectly through N transfer to companion crops or pasture grasses. The latter leads to soil fertility improvement and subsequently enhanced crop/grass and animal productivity. N transfer is closely linked to soil organic matter (SOM) dynamics and the decomposition of leaf litter and legume roots (Cadisch et al., 1998). Litter production of D. ovalifolium can be as high as 6-7 t ha^-1 a^-1 (Ara et al., 1991). Costa (1995) and Gonçalves & Costa (1994) observed a transfer of 43 and 65% of the BNF derived N (70 and 154 kg ha^-1 a^-1) to Pennisetum purpureum and Brachiaria humidicola, respectively. At Itaguí, Brazil, N fixation by D. ovalifolium was rather low (46-68 kg N ha^-1 a^-1) (Viera-Vargas et al., 1995), in contrast to Porto Velho, Brazil, where 137 kg N ha^-1 a^-1 were found, but of which only 28% were transferred (Costa, 1993). An even lower transfer rate of 4% was found by Gonçalves et al. (1992) also in Porto Velho.

Further reading

Entire bibliography on Desmodium ovalifolium

Links

• Effect of boron on forage quality of D. ovalifolium
• D. ovalifolium references
• Short CIAT report
• Ush fallow research with D. ovalifolium
• D. ovalifolium under coconuts in Vanuatu
• Early Growth of Young Oil Palm Under Different Leguminous Cover Crops
• Potential nitrogen fixation of five Desmodium heterocarpon subsp. ovalifolium genotypes

Key References

Other References