Centrosema pubescens Benth.

Leguminosae

Common names

Centro (Australia), jetirana (Argentina, Brazil), bejuco de chivo (Colombia), campanilla (Colombia), butterfly pea.

Note: It has recently been suggested (see CIAT) that the correct name for the species agronomically known as C. pubescens is Centrosema molle Mart. ex Benth. Consequently, all material of the species agronomically hitherto known as C. pubescens (except cv. Belalto) should be called Centrosema molle Mart. ex Benth., and all material of the species agronomically hitherto known as C. schiedeanum should be called Centrosema pubescens Benth. The change would apply, for example, to all 1208 accessions of C. pubescens and 49 accessions of C. schiedeanum registered in the World Catalogue of Centrosema Germplasm (Schultze-Kraft et al. 1989). Schultze-Kraft R; Williams RJ; Coradin L; Lazier JR; Kretschmer, Jr, AE. 1989.
World catalogue of Centrosema germplasm. Centro Internacional de Agricultura Tropical (CIAT); International Board for Plant Genetic Resources (IBPGR), Cali, Colombia].  

Vigorous, trailing, twining and climbing perennial herb; in pure stands forms a compact dense cover 40 to 45 cm high in four to eight months from sowing. Very leafy; the slightly hairy stems do not become woody for at least 18 months . Leaves trifoliate; leaflets dark green elliptic or ovate-elliptic, obtuse or shortly obtusely acuminate, about 4 x 3.5 cm, slightly hairy, especially on the lower surface. Stipules long, persistent. Flowers large and showy, borne in axillary racemes. Each flower has two striate bracteoles. Flowers bright or pale lilac on either side of a median greenish-yellow band with numerous dark violet stripes or blotches. Pod linear with prominent margins 7.5 to 15 cm, long, flat, thick, straight or slightly twisted, acuminate, dark brown when ripe, containing up to 20 seeds: septa between seeds. Seeds shortly oblong to squarish with rounded corners, 4 to 5 x 3 to 4 mm, brownish-black, mottled darker blotches with lighter coloured halo. Wilson and Lansbury (1958) stated it had a shallow root system averaging 30 cm in depth with two taproots to each 900 cm2, growing in a granitic sandy loam in Ghana. Monteiro and Aronovich (1966) gave some anatomical details of its vegetative organs.

Native to tropical South America. Introduced to the Malay Peninsula and Indonesia as a cover crop, probably during the nineteenth century. Now widely grown in the tropics, 50 species occur naturally in South America.

Widespread below 600 m (Crowder, 1960). Atkinson (1970) records Centrosema occurring from sea level to 915 m.

Prefers the wet tropics with a rainfall in excess of 1 750 mm or irrigation, but grows in areas receiving 750 mm or more. It does well at Serere, Uganda, which receives 1 325 mm a year with a five-month dry period (Horrell, 1958). Wilson and Lansbury (1958) state that it requires a minimum of 1 000 mm/year of "twin peak" rainfall in Ghana and gives luxuriant growth when rainfall exceeds 1 750 mm.

Deep-rooted and so is fairly drought-tolerant. Dry-season growth slow (Parbery, 1967a), drops its leaves in a prolonged drought (Stobbs, personal communication). Payne et al. (1955) state that it provides some green feed in the dry season at Sigatoka, Fiji.

Fairly good, better than siratro (Macroptilium atropurpureum) but not as good as puero (Pueraria phaseoloides). Farinas (1966) reported that it survives stagnant water for at least two months in the Philippines.

Will grow on a wide range of soils, from sandy loams to clays. Grows vigorously on alluvial soils and hill soils of Fiji. In Sri Lanka it prefers the heavier clay loams to clays (Santhirasegaram, personal communication). Will nodulate in soils with a pH as low as 4.0 (Andrew, personal communication), but optimum pH lies between 4.9 and 5.5. Rijkebusch (1967) says it needs a lime-rich soil. Döbereiner and Aronovich (1966) used lime to eliminate manganese toxicity in Brazil, resulting in a 65 percent increase in the amount of nitrogen fixed. Lime pelleting increased the nodule numbers as much as liming, but did not eliminate the serious effects of Mn toxicity on nitrogen fixation.

Native Centrosema species are well nodulated throughout Central America, but Bowen and Kennedy (1961) showed that Centrosema pubescens has some specificity toward Rhizobium with heritable variation in nodulating ability. With culture QA522 they obtained 2 096 g dry matter per pot with 16.7 percent crude protein, whereas uninoculated plots yielded only 1 334 g with 10.5 percent protein. The current Rhizobium culture used in Australia is CB1103 (1970). Van Rensburg (1967) showed that centro produced abundant nodules of medium size, sometimes in clusters, most frequently on the secondary roots. Oke (1967a) found that centro developed its first nodules two weeks after germination and the young nodules were the most active in fixing nitrogen.

Fair to good in a fertile environment.

Will establish quite well in roughly prepared seed beds provided fertility requirements are met. Establishes well in ashes after burning forest. Gives its best performance on a well-prepared seed bed.

In small areas, centro can be established with a small "Planet Junior"-type planter; in larger areas it can be drilled or broadcast. When planting with a "Planet Junior"! it is sown in rows 1 m apart and 50 cm apart in the row. Should give a complete cover in four to six months. Under certain circumstances can be Oversown into existing pastures. Risopoulos (1966) introduced it into Imperata cylindrica pastures in Zaire after passage with a "Rome" heavy tandem disc plough. Roberts (personal communication) introduced it into a Brachiaria humidicola pasture in Fiji which had been weakened by army-worm attack, and mown. Seed was scarified, inoculated and sod-seeded into the pasture with 550 kg./ha superphosphate. In northern Queensland, it is sometimes introduced into molasses grass pastures after the grass has subdued the weeds, being sown with guinea grass.éxêé

The quite large seed can be sown to depths of 2.5 to 5 cm without affecting germination. Should be rolled or lightly covered by harrow after seeding.

Sow 3.3 to 4.4 kg./ha, drilled in prior to the rainy season. For green manure it can be sown up to 8 kg./ha (Rijkebusch, 1967) . For broadcasting, increase the seeding rate.

39 600. Percentage of hard seed not more than 75.

Storage under constant damp conditions depresses total viability and increases hard seed content (Wycherley, 1960). Serpa (1966) showed that germination is hindered by impermeability of the seed coat, which is genetically controlled. The following methods can be used to break dormancy: (a) scarify mechanically; (b) immerse in concentrated sulphuric acid 24 or 36N for seven minutes, then thoroughly wash with water (Black, 1968); (c) immerse in hot water at 77°C for 15 min. (Stobbs, 1969b) or in boiling water, adding 1/4 cold water, and soak seed overnight (Grundy, 1959); (d) immerse in warm glycerine at 30°C for two hours (Wycherley, 1960); (e) Osram irradiation for 16 hours or more (Wycherley, 1960); (f) warm to 50°C for up to eight hours (Wycherley, 1960). These methods increase germination from 9 to 16 percent.
Inoculation is necessary. Pelleting is not usually necessary, except in high manganese soils, or to protect Rhizobium. For insect control, use an ant and bean-fly repellent­dust with lindane at 85 g/kg. dry seed. If using ant repellent, use lime-pelleted seed to protect the Rhizobium.

Centro grows well in fertile soils without fertilizer. In poor soils it responds to phosphorus and molybdenum and sometimes to magnesium. It is not demanding of potassium. In Brazil, dolomite containing 10 percent magnesium is applied three to four months before planting. Parbery (1967a) doubled the yield of centro on Cunnunurra clay and nearly trebled the yield on Cockatoo sand with 100 kg. N/ha.

• Calcium: 

The plant responds to liming. Calcium content of the leaf, stem and nodules was increased by liming to raise the pH from 5.0 to 6.0. The treatment also increased the content of molybdenum and reduced the manganese content (Watson, 1960). Centro gave 52 percent of its maximum yield in the absence of calcium, the maximum yield being at 1 100 kg./ha (Andrew and Norris, 1961). At maximum growth the Ca content 1.4 percent (Wilson and Lansbury, 1 958), or 1.5 percent according to Andrew and Norris (1961). The Ca uptake is 1.7 to 2.0 percent of the dry matter (Andrew and Hegarty, 1969) . Kannegieter (1966) recorded 1.66 percent of CaO in the dry matter of centro in Ghana.

• Phosphorus: 

The critical level for P in the dry matter of the leaves at the immediate preflowering stage is 0.16 percent (Andrew and Robins, 1969a). Grof (1966) showed a straight-line increase in the dry weight of Centrosema seedlings up to at least 11 0 kg. P2O5 per hectare.

• Potassium: 

Hamilton and Pillay (1941) found that there were no visible symptoms of K deficiency in centro with a content of 1.35 to 1.88 percent potash in the dry matter. Andrew and Pieters (1970a) showed no deficiency at 1.18 percent. Andrew and Robins (1969) showed that deficiency symptoms occur at a K content below 0.85 percent of the dry matter, and Andrew and Pieters (1970a) have reproduced deficiency symptoms, in colour, of plants containing 0.35 percent K in the dry matter. Deficiency symptoms commenced as a general chlorosis on the mid- to lower positioned leaves of the plant. Chlorosis was interveinal in the early stages, but later full chlorosis occurred with no differentiation of veins except that the base portion of the leaf remained dark green in colour. At this stage there was an associated necrotic spotting and in extreme cases marginal necrosis of the leaflet tips and edges.
Associated with the severe form there was a downward curling of the leaflets from the tips and the surfaces became puckered; the interveinal areas became raised above the veins, which resulted in prominent venation of the underside of the leaflets.
In this species, very little leaf abscission occurred and, furthermore, symptoms were spread over much more of the plant. The parts of the plant remaining visually unaffected were the growing points and a few of the older leaves at its base (Andrew and Pieters, 1970a).

• Copper: 

Centro gave 45 percent of its maximum yield in the absence of copper when grown in soil culture. The copper content of the seed is 16.5 ppm (Andrew and Thorne, 1962). The initial effects of copper deficiency were reduced plant growth and a general paleness in colour of the younger leaves and slight interveinal chlorosis. Following this the newly expanded leaves showed marginal necrosis, with necrotic tissue curling upward to give a dished or rolled effect. There was also dieback of the growing tip, usually including the newest expanding leaf, but no subsequent axillary growth occurred. Older stem growth and root growth appeared normal.

• Molybdenum: 

Application of molybdate to centro had a marked effect on the molybdenum content of leaf, stem and nodules, and the nitrogen content was also increased. Liming to pH 7.0 released high levels of molybdenum from the soil (Watson, 1960). Plants with molybdenum deficiency have small white nodules (Andrew, personal communication).

Grows well with Panicum maximum, Hyparrhenia rufa, Melinis minutiflora, Chloris gayana, Pennisetum purpureum, Paspalum dilatatum; less successfully with Brachiaria Utica and Digitaria decumbens. Sometimes planted with Calopogonium mucunoides and Pueraria phaseoloides to give a quick cover, but Calopogonium may dominate (Wilson and Lansbury, 1958) in the first year. Panicum maximum/Centrosema pubescens pastures were the most successful in Fiji (Payne, 1955).

Bailey (1970) states that from five weeks of age to vining, 2,4-D can be used at 0.825 kg./ha acid equivalent, but some check to growth and loss of plants must be expected. Surviving plants recover and growth on sprayed areas is much the same as on those unsprayed after 12 months. This treatment should only be used when alternative methods of control are unsatisfactory. 2,4-DB is not recommended.
Established centro is checked only slightly by 2,4-D at 1.65 kg. of acid equivalent per hectare during the winter months, when growth rate is slow. If the plant is flowering at the time of spraying, germination of the seed set may be lowered by about 15 percent, but this would be of little consequence in a pasture. During the warm spring and summer months, established centro in active growth may be more susceptible to 2,4-D than during the cooler months. Diquat can be used at 285 g/hectare of cation from five weeks of age. Some leaf fall and slight loss of centro numbers can be expected but surviving plants recover rapidly.

Rather slow to establish (but faster than Neonotonia wightii). Oke (1967a) found that centro showed a retardation of growth during the third week from emergence due to exhaustion of seed reserves before effective nodulation took place.

When established, quite a vigorous grower, better in the second year. Produces abundant foliage without woody stems up to about 18 months of age with a cover from 50 to 70 cm thick (Crowder, 1960). Bowen (1959) studied the growth rate of Centrosema pubescens at Ormiston in south Queensland and at Ayr in north Queensland. Growth rate increased to a maximum at the time of runner formation (90 days at both sites), after which this level was maintained until early midwinter, when a decline to zero occurred until regrowth commenced in spring. Decline and cessation of growth were associated at both sites with minimum screen temperatures of approximately 12.8°C. Growth continued at Ayr for approximately five weeks longer than at Ormiston.

At South Johnstone, north Queensland, Centrosema pubescens has fixed 1 00 kg. N/ha in the surface 15 cm of the soil (Bruce, 1967) . Bruce (1965) showed that the total N percentages in the top 7.5 cm of soil under 11- and 16-year-old Panicum maximum/Centrosema pubescens pastures were 0.406 and 0.432 percent respectively, while that under the adjacent rain forest was 0.423 percent, indicating the ability of a legume-based mixed pasture to maintain fertility after clearing the rain forest.
Bruce (1967) recorded that land previously under rain forest which had subsequently been supporting pure guinea grass pasture for 16 years contained 0.272 percent N in the surface 15 cm of soil, while that under a guinea grass/Centrosema pubescens pasture contained 0.346 percent. He also found that an elephant grass (Pennisetum purpureum) / Centrosema pubescens pasture raised the soil nitrogen in the surface 15 cm of soil by 365 kg. N/ha over a three-year period, or at an average rate of 135 kg. N/ha/year. Schofield (1945), after ploughing in an 18-month growth of Centrosema pubescens, found that the soil nitrogen level rose to 71.7 ppm, compared with 34.4 ppm under a bare fallow for that length of time.
Watson (1957b) recorded the excess nitrogen due to fixation by centro to be 235 kg./ha in five months. He also assessed that 1 hectare of Centrosema plants contained the equivalent of 1 303 kg. sulphur of ammonia after four months' growth (Watson 1957a). Horrell and Newhouse (1965) at Serere Uganda, showed that unfertilized grass-legume (Centrosema pubescens + Stylosanthes guianensis) pastures yielded an equivalent amount of dry matter to a grass pasture fertilized with 165 kg. nitrogen per ha. In Uganda, Moore (1962) showed that the nitrogen content of the soil under a Cynodon plectostachyus/Centrosema pubescens pasture was 275 kg. per hectare per year higher than under a pure grass pasture, and the legume raised the nitrogen content of the grass from 1.8 to 2.4 percent. Whitney (1966), in Hawaii, showed that in a volcanic soil Centrosema pubescens fixed 264 kg. nitrogen per hectare in pure stand, and 121 in a mixed grass sward, and transferred 6 to 11 percent of this fixed nitrogen to the associated grass. Nitrogen transfer to the grass occurred only after six months.

Persists well under grazing and has remained in association with Panicum maximum pastures at Innisfail, Queensland, for more than 20 years. In Fiji, pastures of guinea grass and centro have been grazed successfully at bimonthly intervals (Payne et al., 1955). Crowder (1960) reported that in Colombia it persists under rotational grazing but not under continuous heavy grazing.

It is advisable to allow Centrosema pastures to become well established in the first year before grazing. Slashing or roller-chopping and spot treatment with herbicides may be required to subdue weed growth. Thereafter, the aim should be to keep the legume growing vigorously in association with the grass, keeping the grass in check to avoid dominance. Grass is best kept at a height of 37.5 to 45 cm.

Responds well if established; regenerates from seed after a burn.

Self-pollinated, selections breed completely true.

Payne et al. (1955) recorded an average yield of 4 950 kg. DM/ha/year over the three years 1950-52 at Sigatoka, Fiji. Wilson and Lansbury (1958) recorded an annual mean yield of above-ground green matter at 13.5 tonnes/ha. The aerial portion contributed 47 percent of the total plant weight (leaflets 25.8 percent, vines, petioles etc. 21. 1 percent), the roots and stolons 53 percent (primary roots 26.1 percent, adventitious roots 10.5 percent, stolons 12.8 percent). In Brazil, an average yield of 40 tonnes of green matter/ha/year has been recorded.

Makes good hay if cut after flowering and before seeding. Hay has been made successfully in Costa Rica (Horrell, personal communication), in Ghana (Wilson and Lansbury, 1958) and in Colombia (Crowder, 1960). The mown plant is allowed to cure for a day or two and is then baled. Leaf shattering is a problem. Satisfactory silage has been made with a mixture of guinea grass and centro in north Queensland (Teitzel, 1969b) .

Has proved valuable in Uganda for standover feed in association with Hyparrhenia rufa when the dry season starts abruptly (Horrell, 1958).

Valuable for grazing and as a hay crop.
Digestibility of dry matter (Ghana) 53.5 percent; organic matter 53.2 percent; crude protein 62.4 percent; ether extract 44.3 percent; N-free extract 61.3 percent; crude fibre 39.5 percent and ash 56.0 percent. Otero (1952) listed the digestibility of the protein at 65.7 percent and of the fibre at 51.9 percent. Reyes (1955) recorded 35 percent digestibility of the crude fibre in feeding trials in the Philippines. Starch equivalent in Ghana was 10 percent, digestible crude protein 3.2 percent and nutritive ratio 1:3; for the Philippines, Reyes assessed figures of 8.8, 3.2 and 1:2.14 respectively. Miller and Rains (1963) showed that the intake of centro dry matter was low and that the digestibility of the organic matter was low, particularly when immature, relative to that of protein.

None recorded. Seed fed to mice had no ill effects (Bindon and Lamond, 1966).

Seed is harvested by hand in many tropical countries. It is difficult to harvest mechanically because of uneven ripening (Crowder, 1960). Best grown on trellises or fence-lines to facilitate hand-picking (Wilson and Lansbury, 1958). In north Queensland (Gude, 1959), centro seed has been harvested mechanically from late winter to early spring from a molasses grass/centro pasture by two methods: (a) the pasture is mown and windrowed soon after the first seed pods open. The hay is then cured for a few days in the field and when sufficiently dry for threshing it is collected by a pick-up machine harvester; (b) direct heading at a height of 20 to 40 cm. In this case the harvested seed has to be dried on a tarpaulin or concrete floor before cleaning and storage.

At Serere, Uganda, 220 to 275 kg./ha are usual. Wilson and Lansbury (1958) recorded about 1 250 seeds/m2 collected on a fence-line. Average yield from direct mechanical heading in north Queensland is 140 kg./ha.

There is only one registered cultivar of Centrosema pubescens, cv. Belalto, although there are many native ecotypes which include both early- and late-flowering types. 'Belalto' is readily distinguishable from the "common centro" used in Queensland for a number of years by its purple or brown young leaflets and deep mauve to white flowers. It is a vigorous grower in cool conditions and can resist attacks of leaf spot disease and red spider. It also resists weed invasion. 'Belalto' is expected to replace common centro.
Rijkebusch (1967) reports two varieties (M20 and M301) at the Sisal Research Station, Mlingano, Tanzania.GO

Attacked by Cercospora leaf spot at Koronovira, Fiji, at Gualaca, Panama, and at Innisfail, Queensland­all areas receiving over 2 500 mm rain a year. The attack can be serious. Aschochyta blight affects it in Guatemala. A fungus sometimes affects the pods in Colombia.

Ability to persist under high rainfall and into the dry season as carryover feed; mixes well with grasses; gives good live-weight gains in the dry season as carryover feed; mixes well with grasses; gives good live-weight gains in the dry season; good seed production, easily obtained; good green manure crop in rubber, coconut and oil-palm plantations.

Only moderately stoloniferous cool season dormancy; requires a higher fertility level than stylo; low consumption in the wet season; intolerant of low grazing height.

Wilson and Lansbury (1958) reported that West African Dwarf Forest sheep grazing Centrosema gained 6.7 kg. per head in 11 months. Daily dry-matter intake averaged 3 percent of the body weight. As a sole diet, it was thought to be too rich in protein, and a supplement of cassava root was suggested for better performance in Ghana.
An irrigated guinea grass/centro pasture at Ayr, north Queensland, under 41 weeks' grazing with shorthorn beef cattle gave a daily live-weight gain of 0.68 kg./head, and a Para grass/centro pasture 0.74 kg. per head. From January to March, weight gains fell owing to high day temperatures and high humidity (Allen and Cowdry, 1961a).
In Uganda, the daily live-weight gain by White Fulani cattle grazing Cynodon plectostachyus/Centrosema pastures was 0.37 kg. compared with 0.31 kg. from pure grass (Moore, 1962).
The inclusion of Centrosema in a guinea grass pasture at South Johnstone, north Queensland, increased live-weight gain from 440 to 610 kg./ha on pure grass swards (Grof, 1966).
Stobbs (1966) obtained a mean of 495 kg. live-weight gain per hectare per year over a two-year period with cattle grazing a Hyparrhenia rufa/ Centrosema pubescens sward at Serere, Uganda (lat. 1°32'N, rainfall 1 360 mm/ year) . The effect of introducing the legumes Stylosanthes guianensis and Centrosema pubescens together with a Hyparrhenia rufa, Panicum maximum and Chloris gayana mixed pasture with and without added nitrogen at Serere is shown in Figure 48.
Stobbs and Joblin (1966b) found that Centrosema pubescens increased the production of Hyparrhenia rufa by 18 percent at Serere, Uganda, with an advantage of 11 percent in the second year of grazing and 26 percent in the third year. Hyparrhenia/centro pastures produced a live-weight gain of 815 kg. per ha over 21 months (465 kg./ha/year) and Panicum maximum/centro pastures produced 392 kg./ha/year. The contribution of centro to the dry-season weight gain was 28 percent or 10 kg. per month, and during the wet season 15 percent or 6.4 kg. per month. Centro produced palatable green material during the three-month dry season.

Allen and Cowdry (1961a, b).

2n = 20.

Low, and is severely damaged. Needs the protection of its crown and lower mature stems. Once established has better survival value.

Probably about 22°N and S latitude.

Will grow well with grasses of medium height. Seedling growth slow under shady conditions, but mature plants tolerate shade.

Reasonably competitive, but slashing helps.

Meloidae beetles and thrips in Uganda (Horrell, 1958) and red spider (Tetranychus sp.) at Innisfail in north Queensland attack the leaves (Grof, personal communication).

Summer; prefers moist conditions with a maximum of 25.6°C. Dormant during the winter. Ludlow and Wilson (1970) recorded only 5.3 percent dry-matter production when grown at 20°C compared with the growth at 30°C. Bowen (1959a, b) showed growth ceased at 12.8°C .

Centro tolerates fairly high levels of manganese in the soil. Döbereiner and Aronovich (1966) believe that 300 ppm Mn in the foliage is harmful, but Andrew and Hegarty (1969) assessed the "toxicity threshold value" (the Mn concentration in the dry matter when yield is 5 percent lower than the maximum) for Centrosema pubescens at 1 600 ppm.
In young plants of this species, the dominant effect of manganese toxicity was a general interveinal chlorosis of the younger leaves and shoots, very like the effect of iron deficiency. However, the young expanding leaves, while being severely chlorotic, also had a bronze colour, the result of numerous minute, rusty coloured spots over the entire leaf surface. The young, fully expanded leaves showed intense interveinal chlorosis, and the margins of the leaflets, particularly toward the tip, became necrotic and curved outward away from the midrib, giving each leaflet a convex dish effect. There was also slight epinastic curvature of the petiolules (Andrew and Pieters, 1970b)­ the upper surface growing faster than the lower.

Fairly palatable.

Germination 60 percent, less than 10 percent hard seed, purity 97.5 percent. The seed is germinated at 25°C (Queensland).