Previous Page Table of Contents Next Page

3.6 Seed Production of Forage Tree Legumes

R.C. Gutteridge and W.W. Stür

Principles of Reproductive Development
Practical Aspects of Seed Production


Apart from Leucaena leucocephala very little definitive information is available on the seed production of forage tree legumes. Leucaena leucocephala is the only species for which seed has been produced in commercial quantities but production is declining in the wake of the psyllid problem. Small quantities of seed of species such as Gliricidia sepium, Calliandra calothyrsus and Sesbania sesban are available from various institutions and organisations mainly for research purposes but reliable sources of larger amounts of seed are virtually nonexistent.

Little detailed research on seed production of forage tree legumes has been carried out and this section relies on principles applicable to other plants as well as on observations.

Principles of Reproductive Development

Although tree legumes are generally long-lived, most are dependent on seed production for long-term survival and each species has had to evolve a reliable system of seed production.

Climate has a major influence on the evolution of seed production characteristics since the timing of flowering, pollination and seed development is crucial for successful seed production.

Reproductive development can be divided into several phases. These include a period of obligate vegetative growth (juvenility), floral initiation, anthesis, pollination and seed development.


Plants first pass through a juvenile phase before they are able to move into a reproductive phase. The length of the juvenile phase varies from species to species and can be as long as several years. It appears to be related to the length of the life cycle of the species and the climate in which the plant has evolved. For example, seedlings of annual grasses from arid environments, where the growing season is short, can initiate floral development within days of emergence (Humphreys 1981). Once plants have passed through their juvenile phase, floral initiation can occur whenever conditions are favourable.

In Java, Calliandra calothyrsus has been reported to flower 4-6 months after planting (NAS 1983). In Brisbane, Australia, Sesbania sesban flowered 3 months after emergence (Sedi and Humphreys 1992).

Floral initiation

Floral initiation is usually controlled by environmental factors such as daylength, temperature and water and nutrient availability. The timing of floral initiation is related to the climate in which the species has evolved. For example, Gliricidia sepium originates from Mexico and Central America (latitude 7-25°N) in a sub-humid climate with 5 months' dry season from December to April (Hughes 1987). In this environment, trees are deciduous, losing their leaves early in the dry season in December/January, and flower (without leaves) between January and March. Seeds mature 40-55 days after flowering before the onset of the first rains in May. Hughes (1987) stated that gliricidia produces a good seed crop in most years and that the timing of seed production is highly predictable in its areas of natural distribution. Floral initiation may therefore be related to:

· short days (shortest day is approximately 11 h at latitude 15°N),
· dry season moisture stress, or
· deciduousness, which in turn is often related to photoperiod.

Although temperature varies little near sea level at latitude 15°N, flowering is delayed by low temperatures experienced at higher altitudes. Hughes (1987) reported that in Guatemala, seed of gliricidia matures in February at sea level but 2 months later at a nearby high altitude (950 m) site. It is likely that all four factors interact in determining the time of floral initiation in gliricidia.

When grown at other similar locations around the world, flowering and seed production follow a similar pattern. In Timor, Indonesia (latitude 10°S), gliricidia sheds its leaves at the beginning of the dry season in July and flowers in August (= shortest days, dry season and deciduousness) (J. Nulik, personal communication). At latitude 10°N in the Philippines, gliricidia sheds its leaves at the beginning of the dry season in November and flowers in December (= shortest days, dry season and deciduousness) (A. Castillo, personal communication). In Brisbane, Australia (latitude 27°S), gliricidia drops its leaves in July (shortest days and low temperatures) and flowering is delayed until October (Gutteridge and MacArthur 1988). If grown at locations where the environmental conditions are markedly different gliricidia will not produce a high seed yield. For example, Glover (1989) reported that in some parts of Asia and the south Pacific, gliricidia flowers in the wet season and seed production is poor.

In general, floral initiation of temperate and subtropical species is strongly related to photoperiod (varied somewhat by temperature), while this factor is generally only weakly expressed (quantitatively rather than qualitatively) in tropical species. Tropical species are usually able to initiate floral development throughout the year and sometimes reproductive development is enhanced by a change from wet to dry conditions. Some species, with a short day requirement, do not or only sparsely flower near the equator, but can be induced to flower more profusely by low temperatures (high altitude) or by moisture stress (Humphreys and Riveros 1986).

In leucaena, flowering can occur at any time during the year whenever growing conditions are favourable but will increase under moisture stress and also with the onset of shorter days in the subtropics. Calliandra produces most seed in the June to September dry season in Java; however, it is capable of flowering throughout the year (NAS 1983).

Anthesis, pollination and seed development

The mode of reproduction or breeding system of tree legume species varies from largely self-fertilised and self-compatible as in Leucaena leucocephala (Hutton and Gray 1959) to largely outcrossing. Examples of outcrossing species are gliricidia (Aken'Ova and Sumberg 1986), Erythrina spp. (Neill 1988) and Prosopis spp. (Arroyo 1981).

In leucaena, anthesis occurs early in the morning and pollen falls directly on to the stigmas; this results in a high degree of self-pollination, although cross-pollination can occur (Hutton and Gray 1959). Other species of the genus Leucaena such as the diploid form of L. diversifolia and L. pallida are self-incompatible and thus highly cross-pollinated (Brewbaker 1983).

In cross-pollinated species, the success of pollination depends on the presence of suitable pollinators and favourable weather during anthesis. Carpenter bees (Xylocopa sp.) have been observed visiting gliricidia flowers both in its native range and in areas where it is cultivated, and Janzen (1983) suggested that these bees are the primary pollinators. Calliandra is pollinated in its native range by bats of the genus Glossophaga and by large hawkmoths. Sesbania spp. are pollinated by bees, except for large-flowered species such as S. grandiflora which appear to be pollinated by birds (Brewbaker 1990).

Seed development from pollination to seed maturity takes 40-55 days in gliricidia (Hughes 1987), approximately 60 days in calliandra (NAS 1983) and 40 days in Sesbania cannabina in good growing conditions (Sedi and Humphreys 1992).

Once seeds are mature, the pods containing the seed shatter and seeds are dispersed. This dispersal mechanism has implications for harvesting, since pods on a tree ripen unevenly and hence have to be harvested individually.

Seed pods and seeds vary greatly in size and shape between species. Seed size ranges from 55,000 to 80,000 seeds per kg in Sesbania sesban to 21,000 to 28,000 seeds per kg in leucaena, 14,000 seeds per kg in calliandra and 4,70011,000 seeds per kg in gliricidia.

Practical Aspects of Seed Production

Site selection

Selecting the 'right' site for seed production is of utmost importance. Apart from providing good growing conditions for the tree legume species, the timing of floral initiation, anthesis and seed development is crucial if maximum seed yields are to be obtained. Adverse climatic conditions during pollination or seed development can result in very low seed set and yield. For example, Atta-Krah (1987) reported that early rains during flowering led to heavy flower drop and low seed set in gliricidia. Seed production problems in calliandra have been widely reported in Africa, particularly Kenya (D.J. Boland, unpublished data). The reason for this is unclear but it may be related to the lack of suitable pollinators or climatic incompatibility with heavy rains preventing pollen movement during flowering.

Areas with reliable and distinct wet and dry seasons, where the tree legume produces seed during the dry season, are ideal for seed production because of the favourable warm, dry conditions during flowering, pollination and seed development. A fertile, well drained soil will favour good seed production.

Seed crops of largely outcrossing, insect-pollinated species such as gliricidia need to be isolated from other trees of the same or related species to prevent cross-pollination. Such species need to be planted in blocks containing at least 30 trees and isolated by at least 200 m. A border row should be established around the block and seed should not be collected from this row. Largely self-fertilised species such as Leucaena need only be separated from related species by a few metres to prevent contamination.

Management and harvesting

To obtain maximum seed yields, trees need to have a good framework to maximise potential floral sites. Cutting gliricidia to 0.5 m after a seed harvest reduced flowering and seed yield in the following 2 years relative to an uncut control (Atta-Krah 1987). Similarly, Mohatkar and Relwani (1985) reported that seed yield of leucaena was higher when cut at a height of 1.2 m rather than 0.6 m and that seed yield decreased with higher tree density. While this demonstrated a positive relationship between tree size and seed yield for individual trees, practical considerations for harvest of large trees need to be taken into account. Low seed yield per tree may be compensated by a higher planting density. Hare (1985) suggested that leucaena seed crops should be sown in rows 2 m apart with 0.2 m between plants in the row. The trees should be cut back to 0.5 m height at the start of each wet season to encourage branching and facilitate hand harvesting.

The potential seed yield of tree legume species is very high. For example, seed yields of up to 3,000 kg/ha were recorded for leucaena from a small seed production orchard in Hawaii (NFTA 1985). However, the actual harvested yield is much lower due to harvesting problems; pods are often at an inaccessible height and pods ripen unevenly which means that only a small fraction of the potential seed yield is harvestable at any one time. Repeated hand harvesting can overcome this problem to some extent. In Australia, mechanical harvesting of Leucaena has been achieved using overhead booms or arms supported on a frame which knocks the ripe pods into a trailer. A mobile thresher and cleaner can be attached to take the seed from the collecting trailer (J. Wildin, personal communication). In 1986/87, over 5 t of L. leucocephala seed were produced for commercial sale in Queensland (QDPI statistical records), but this declined to less than 3 t in 1987/88 following the arrival of the psyllid insect and consequent reduction in demand for seed.

Gliricidia regularly yielded over 25 kg/ha of seed per year in a 5 ha plantation near Pen Fui airfield in Timor, Indonesia (J. Nulik, personal communication). In West Africa, Sumberg (1985) reported seed yields of gliricidia up to 89 g per tree per year, equivalent to approximately 37 kg/ha at the spacing used. Seed yield was closely related to the number of set racemes per tree. Seed collection from most of the perennial sesbanias is easy and large quantities of seed can be rapidly hand harvested and processed. In Hawaii, Evans and Rotar (1987) obtained seed yields of approximately 1.5 t/ha from S. sesban, although R.C. Gutteridge (unpublished data) obtained lower yields of approximately 250 kg/ha from a small area of S. sesban var. Nubica at Mt Cotton, southeast Queensland. Harvested seed needs to be dried in the shade to a moisture content below 10% for safe storage. Fast sun-drying may damage seeds and reduce germination percentage and storage life.

Seed storage

In general, the seed of tree legumes with a hard seed coat (e.g. Leucaena spp. and most Sesbania spp.) can be stored for reasonably long periods without special treatment. Lulandala (1981) obtained 100% germination percentage of leucaena seed after 11 months' storage at 25°C. Cobbina et al. (1990) reported that the germination of Leucaena seed stored at either room temperature, in a deep freeze or in dry soil in a glasshouse for 12 months was not significantly different from that at the start of the storage period.

However, seed of species without a hard seed coat such as gliricidia, S. grandiflora and to a lesser extent calliandra requires specialised storage conditions if seed viability is to be maintained. Ideally seed of these species should be stored in sealed containers at a moisture content of less than 10% and at a temperature of 4°C or less. Calliandra seed retained viability when stored in a refrigerator (4°C) for 2.5 years but viability was reduced by 15% when seed was stored at room temperature for one year (NAS 1983).

Thomson and Evans (1990) suggested that seed storage life will be prolonged


· ensuring that the seed collected is fully mature, thoroughly dried (<10% moisture) and clean,
· dusting with insecticides to kill insects,
· storing in airtight containers in an atmosphere of CO2, and
· storing at low temperatures (< 4°C) and low humidity.


Although seed production of forage tree legumes has not been widely studied, it appears that it should not be unduly difficult to produce reasonable quantities of good quality seed provided suitable locations are selected.

There is scope for the selection or breeding of varieties of tree legumes with higher seed yields with improved synchronisation of seed set, reduced pod shattering on drying and resistance to pod pest infestation. However, higher seed yields should not be achieved at the expense of vegetative yield and the ultimate aim of all improvement programmes must be higher yields of edible forage.


Aken'Ova, M.E. and Sumberg, J.E. (1986) Observations of the pollination system of Gliricidia sepium. Nitrogen Fixing Tree Research Reports 4, 29-30.

Arroyo, M.T.K. (1981) Breeding systems and pollination biology in Leguminosae. In: Polhill, R.M. and Raven, P.H. (eds), Advances in Legume Systematics (Part 2). Royal Botanic Gardens, Kew, pp. 723-769.

Atta-Krah, A.K. (1987) Flowering and seed production of Gliricidia sepium. In: Withington, D., Glover, N. and Brewbaker, J.L. (eds), Gliricidia sepium (Jacq.) Walp.: Management and Improvement. Proceedings of a workshop held at CATIE, Turrialba, Costa Rica. NFTA Special Publication 87-01, Hawaii, pp. 142-145.

Brewbaker, J.L. (1983) Systematics, self-incompatibility, breeding systems and genetic improvement of Leucaena species. In: Leucaena Research in the Asian-Pacific Region. Proceedings of a workshop held in Singapore, IDRC Ottawa, pp. 17-22.

Brewbaker, J.L. (1990) Breeding systems and genetic improvement of perennial sesbanias. In: Macklin, B. and Evans, D.O. (eds), Perennial Sesbania Species in Agroforestry Systems. Proceedings of a workshop held in Nairobi, Kenya, NFTA Special Publication 90-01, Hawaii, pp. 39-44.

Cobbina, J., Kolawole, G.L. and Atta-Krah, A.N. (1990) Leucaena and gliricidia seed viability and germination as influenced by storage conditions. Leucaena Research Reports 11, 91-93.

Evans, D.O. and Rotar, P.P. (1987) Sesbania in Agriculture. Westview Tropical Agriculture Series No. 8. Westview Press, Boulder, Colorado, USA, 192 pp.

Glover, N. (1989) (ed.) Gliricidia Production and Use. NFTA, Hawaii, 44 pp.

Gutteridge, R.C. and MacArthur, S. (1988) Productivity of Gliricidia sepium in a subtropical environment. Tropical Agriculture (Trinidad) 65, 275-276.

Hare, M.D. (1985) Tropical Pasture Seed Production for Village Farmers in Southeast Asia. Ministry of Foreign Affairs, Wellington, New Zealand, 44 pp.

Hughes, C.E. (1987) Biological considerations in designing a seed collection strategy for Gliricidia sepium (Jacq.) Walp. In: Withington, D., Glover, N. and Brewbaker, J.L. (eds), Gliricidia sepium (Jacq.) Walp.: Management and Improvement. Proceedings of a workshop held at CATIE, Turrialba, Costa Rica. NFTA Special Publication 8701, Hawaii, pp. 174-184.

Humphreys, L.R. (1981) Environmental Adaptation of Tropical Pasture Plants. MacMillan, London and Basingstoke, 261 pp.

Humphreys, L.R. and Riveros, F. (1986) Tropical pasture seed production. FAO Plant Production and Protection Paper 8, FAO, Rome, Italy, 203 pp.

Hutton, E.M. and Gray, S.G. (1959) Problems in adapting Leucaena glauca as a forage for the Australian tropics. Empire Journal of Experimental Agriculture 27, 187-196.

Janzen, D.H. (1983) Gliricidia sepium. In: Janzen, D.H. (ed), Costa Rican Natural History. University of Chicago Press, Chicago, USA.

Lulandala, L.L. (1981) Seed viability, germination and pretreatment of Gliricidia sepium seed orchard. Tropical Agriculture (Trinidad) 62, 17-19.

Mohatkar, L.C. and Relwani, L.L. (1985) Effect of plant population, stubble height and number of cuttings on the growth, seed, forage and firewood production of Leucaena K8. Leucaena Research Reports 6, 40-41.

NAS (1983) Calliandra: a Versatile Small Tree for the Humid Tropics. National Academy Press, Washington DC., USA, 57 pp.

Neill, D. (1988) Experimental studies on species relationships of Erythrina (Leguminosae: Papilionoideae). Annals of the Missouri Botanical Gardens 74, 27-41.

NFTA (1985) Leucaena Forage Production and Use. Nitrogen Fixing Tree Association, Hawaii, USA, 37 pp.

Sedi, Y. and Humphreys, L.R. (1992) Temperature, daylength and the flowering and seed production of Sesbania sesban and S. cannabina. Tropical Grasslands 26, 100-110.

Sumberg, J.E. (1985) Note on flowering and seed production in a young Gliricidia sepium seed orchard. Tropical Agriculture (Trinidad) 62, 17-19.

Thomson, L. and Evans, D.O. (1990) Seed production of perennial Sesbania species. In: Evans, D.O. and Macklin, B. (eds), Perennial Sesbania Production and Use. NFTA, Hawaii, USA, pp. 29-32.

Previous Page Top of Page Next Page