(introduction...)

BNF Technology for Tropical Agriculture

Graham, P. H. and S.C. Harris, Eds.
CIAT, AA 67-13, Cali, Colombia 768 pp. 1982

J.L. Brewbaker, R. Van Den Beldt and K. MacDicken¹

¹ Dept of Horticulture University of Hawaii. Honolulu. 96822 USA.

Summary

Nitrogen (N₂) fixing trees are discussed with special attention to their use as fuelwood, forage or green manure in the tropics. Severe deforestation is viewed as leading to a "balding of the tropics" that could jeopardize the genetic resources of many legume trees. Increasing fuel and fertilizer costs mandate the planting and husbandry of tropical fuelwood and green manure tree crops. High population densities giving maximal annual biomass yields, and the use of trees with little concern about conformity or beauty, will provide attractive targets for breeder and/or agronomist/silviculturist.

N₂-fixing genera with special values as fuelwoods, forages, green manures or nurse trees, ornamentals, and as timber are listed. Characteristics are given for 18 fast-growing N₂-fixing trees in current University of Hawaii network trials.

The balding of the tropics

It is traditional for man to plant and grow his food, but not to grow the wood with which to cook it. In the world of 1900 AD, the hunting and collection of fuelwood from native forests presented little challenge. There were only 1.6 billion people in the world, and approximately seven billion ha of forests. In the world of 2000 AD, however, the challenge of finding fuelwood will be awesome (Food and Agriculture Organization (FAO), 1980). A world population of 6.4 billion is predicted for 2000 AD, with only 3.0 billion ha of remnant forests (down from 4.8 trillion in 1950).

The "people vs. trees" problem is greatly exacerbated in the tropics, where most countries have doubled their human populations in only the past three decades, while cutting their forest lands by half. Forest depletion figures for developing countries arc startling (see Table 1). Forest areas with closed canopy (including growing stock) that totaled 1,270 million ha in 1978 arc predicted to drop to 760 million ha by the end of this century (Barney, 1978). The ramifications of this loss arc staggering, but include possible effects on atmospheric carbon dioxide arid world climate (Woodwell, 1978). In contrast, only a slight loss is anticipated from the 1,620 million ha of closed forests in developed countries. This may be recalled as the century when Planet Earth grew a giant held ring around its equator. The planting and husbandry of fuelwood in the tropics is clearly mandated for the future.

TABLE 1: World forest resources as totals of closed forest area and stock growing (adapted from Barney, 1978).

Region	Total closed forest area (106 ha)
	1978	20001
Tropics
Latin America	640	380
Africa	230	180
Asia/Pacific	400	200
Total	1270	760
Temperate2	1620	1610

1 By comparison, total world forests in 1950 exceeded 5000 million ha.
² North America, Europe, USSR, Japan, Australia, New Zealand.

It has also been traditional for man in the tropics either to allow nature to repair the soil losses to agriculture by the following of land for 15-20 years, or to use inorganic fertilizer. The slash and burn tradition can no longer continue into the 21st century, as the forest depletion and man's population pressure simply obviate it. Neither can inorganic fertilizers be an economic option except for the limited. wealthy fraction of farmers. Thus, the planting and husbandry of green manure crops also becomes a mandate for the future.

The majority of tropical legumes are woody perennials, many of which are troth energy producing and nitrogen (N₂) fixing. It may he asserted that the health of many tropical forests relies initially on leguminous trees for N₂ fixation. Wild populations of native or aggressive introduced leguminous trees are increasingly valued as fuelwoods (National Academy of Science (NAS), 1980) and to a lesser extent as green manure trees (NAS, 1979). Notable among these are tile mimosoids, a subfamily of legumes that includes about 2800 species. predominantly tropical trees and shrubs.

Maximization of biomass/ha per year must be the immediate target for both fuelwood and fertilizer production by trees. Reduced to essentials, the number of carbon and N atoms fixed annually per unit area becomes the goal, with little consideration of tree form or appearance. It is a target more familiar to agronomists than to foresters, and one that gives the plant breeder versatile free rein.

It is safe to predict that fuelwood and fertilizer tree production will be dominated within a few decades by trees that are agriculturally versatile and easily bred and managed. The future improvement of these legumes could, however, be limited by the availability of appropriate germplasm. With the accelerating loss of virgin tropical forests, these native resources are dwindling and are often endangered.

Genetic resources for N2-fixing trees

The Nitrogen -Fixing Tree Association (NFTA), a new international organization that aims to encourage research and communication on leguminous trees, was incorporated in Hawaii in 1981. A primary thrust of the NFTA is to help identify genetic resources and stimulate their careful preservation and expansion. Our present impression is that the genetic resources of N₂-fixing trees are in a tragic state. There are no major international repositories of legume tree germplasm, whether as seed, or in arboreta, and very few tree species have been the subject of botanical expeditions for germplasm collection. Additionally many of the genera of N₂-fixing trees are taxonomically confused, from unknown centers of origin, or from areas that are rapidly becoming treeless. Seeds available for distribution are often of unknown origin. Genetically distinct varieties are available for only a few species, and these are predominantly ornamentals.

The opportunities for exploitation of the genetic diversity in legume trees can be illustrated from studies with Leucaena leucocephala (known also as ipil-ipil, huaxin, guaje, leadtree, lamtoro, koa haole, or kubabul). These have been reviewed by Brewbaker & Hutton (1979) and other authors (NAS, 1977; Brewbaker, 1980). The arboreal leucaenas did not become naturaly dispersed through the tropics, but only a shrub known as the "common-type" or "Hawaiian-type" Leucaena. Though our collection of this heavily flowering shrub includes more than 500 accessions from numerous countries in the tropics, there is little genetic variability. We surmise that all originated from a narrow gene base. The species was dispersed from its native Mexico mainly through Spanish galleons departing from Acapulco and Mazatlan. In this region a highly flowering shrub is the only representative of the species, and it is clearly this one self-pollinated variety that circled the world.

The tree form and other genetic variants of Leucaena occur in southern Mexico and in Central America, a centre of diversity for this tetraploid species (which is an evident hybrid of two other species). The arboreal types were first considered a distinct species by botanists; then came to be known as the "Salvador type." This type first came to Hawaii from Central American seed collectors in the 1930's, and was then widely dispersed in the 1960's as a result of research in Hawaii and in Australia (Brewbaker, 1975). As a source of fuelwood, the Salvador type exceeds the common type by over 100% in wood yield; yet differs by very few genes.

It is virtually certain that genetic gains similar to those in Leucaena await the first plant explorers for species grown solely as C or N. fixers. Since many of these species are outcrossing, unlike Leucaena the identification of genetic superiority will require more care in seed production. However, such species may well afford greater genetic gains-as occurred in poplar and pine - through exploitation of hybrid vigor in controlled crosses or from seed orchard synthetics.

The hazards of endangerment of species are evident in Leucaena. The center of origin of the Salvador type appears to he in the Morazan province of southern Salvador, a region now virtually treeless. Salvador-type leucaenas are now to be found only in the city squares and in backyards, a poor genetic sample of what existed as little as 50 years ago. Leguminous trees are often selectively browsed by feral animals and are, thus, more apt to extinction than many others. Following fire, however, they often regrow with ferocity from the fire-scarified seeds that have long lain dormant in the soil.

Important genera of N2-fixing trees

The 18,000 species of legumes (Family: Leguminosae) include the vast majority of important N₂-fixing trees and shrubs, many of which are in the predominantly woody subfamilies Mimosoideae (2800 spp.) and Caesalpinioideae (2800 spp.). Relatively few of the 12,000 species of Papilionoideae are arboreal, but some of these are of great economic importance. A high proportion of the tested mimosoids (92%) are able to fix N₂, contrasted with the papilionoids (94%) and the caesalpinioids (34%). A few nonleguminous tree genera also fix N₂, notably the temperate genus Alnus and the tropical Casuarina (Stewart, 1967; see p. 427).

Leguminous trees produce some of the outstanding luxury timber of the tropics (NAS, 1979). Notable among these are the papilionaceous genera Dalbergia (rosewood), Perocopsis (African teak), Pterocarpus (narra), and the caesalpinioid genus Intsia (ipil, Moluccan ironwood). Other important timbers include the mimosoids Acacia, Lysiloma, Parkia, and Samanea. Preferred timber species often exceed 30 m in height and are of slow-to-intermediate growth rates. With their high intrinsic value, such trees might wisely be interplanted at wide spacing (e.g., 100/ha) in plantations of fast-growing legumes, as a long-term investment.

The legume trees best known as ornamentals, offering striking displays of color when in flower, are predominantly in the Caesalpinioideae, many of which do not fix N₂. The ornamental legumes include:

Caesalpinioideae: Amherstia, Barklya, Bauhinia, Brownea, Caesalpinia, Cassia, Colvillea, Delonix, Peltophorum, Saraca, and Schotia.

Mimosoideae: Calliandra, Samanea.

Papilionoideae: Butea, Erythrina, Sabinea, Sophora.

Several tree legumes provide valuable gums (Acacia spp.) and the pods of several species are excellent human foods, including:

Caesalpinioidae: Ceratonia (carob), Tamarindus (tamarind).
Mimosoideae: Inga. Parkia.

The following discussions will focus on legume trees with special significance as sources of energy or green manure. As a generalization, most fast-growing legume trees are mimosoids. Genera to be considered in the discussions of energy and green manure are listed below, together with their approximate number of species:

Caesalpinioideae: Acrocarpus (3), Cassia (600), Schizolobium (5)

Mimosoideae: Acacia (600), Albizia (100), Calliandra 9100), Desmanthus (40), Mimosa (450), Parkia (40), Pithecellobium (200), Prosopis (44), Samanea (1).

Papilionoideae: Dalbergia (250), Erythrina (100), Flemingia (35), Gliricidia (10).

Wood and fuelwood

World production of wood in 1975 exceeded 2.5 billion m³ (World Bank, 1978). Less than a century ago, wood was the major energy source for all countries in the world. Today, only 45% of the wood harvested is for fuel, and this is almost entirely in the tropics. Industrial uses of wood (60% in construction, 25% for pulp, 15% for others uses) have increased far more rapidly than total world commodity trade. These uses govern the base price of wood and directly influence both the availability and cost of fuelwood in the tropics. Demand for industrial wood has been increasing at about a doubling rate every 25 years. Demands for fuelwood are also increasing and will soon exceed capacity in regions such as Asia, which has less than 0,18 ha of forest per person at present (Revelle, 1980).

Tree legume species considered of special significance for fuelwood are summarized in Table 2. Species with unusual adaptability to the arid tropics are distinguished. Although many of these species appear to be slow in growth in their native habitats, they are often fast growing under experimental conditions, notably with adequate water. Species of Acacia and Inga provide fuelwoods for tropical highlands, while temperate fuelwoods would also include species of Gleditschia and Robinia.

TABLE 2: Tropical tree legumes of special significance as fuelwood (adapted from NAS, 1980).

Genus	Species adapted to:
	Humid tropics	Arid tropics
Acacia	auriculiformis, mearnsii1	brachystigia, cambagei, cyclops, nilotica, saligna, senegal, seyal, tortilis lebbek
Albizia
Calliandra	calothyrsus
Cassia		siamea
Derris	indica
Gliricidia	sepium
Inga	vera1
Leucaena	leucocephala
Mimosa	scabrella
Pithecellobium		dulce
Prosopis		alba, chilensis, cineraria, juliflora2, pallida, tamarugo
Sesbania	grandiflora

1Highland-adapted species.
²Widely considered an undesirable, thorny pest.

Dendrothermal power plants can be designed to use chips (conventionally) or roundwood. Choice of fuelwood stock is influenced primarily by heat production (combustion value), and by ease of sawing, chipping. anti transportation. Combustion values and wood densities are summarized in Table 3 for the species included in the University of Hawaii studies. Combustion values (given for bone-dry wood) resect wood chemistry, not density, and vary lime for the species listed. These values decrease linearly as wood moisture increases (most fuelwoods contain about 50% moisture at harvest). Specific gravity of species like the fast-growing Albizia falcataria are too low to make commercial fuelwood, due to bulk density problems of transportation and handling for the boiler. On the other hand, some species arc so dense (e.g.. arid-zone Acacia and Prosopis spp.) that they present problems in sawing and chipping. An economic feasibility analysis in Hawaii (Brewbaker, 1980) concluded that giant leucaenas could be grown and harvested profitably as boiler fuel, even with Hawaii's high costs of labor, land, and water. Energy returns from a 1000 ha tree farm, harvested incrementally on a four-year cycle, were calculated to be 28.6 million kwh annually. Wood drying and use of high efficiency boilers could increase this value by 20%.

TABLE 3: Characteristics of N₂-fixing trees in University of Hawaii international network trials (Scale: 1, Good-3, Poor).

Characteristicas	Genus and Species
	a	b	c	d	e	f	g	h	i	j	k	l	m	n	o	p	r	t
Utility for:
Forage	3	3	2	3	2	1	2	3	2	2	1	1	1	3	1	2	3	1
Fuelwood	1	2	1		3	1	1	1	1	2	1	1	1	1	1	3	3	1
Roundwood	3	2	1	3	1	3	1	2	3	1	1	1		3	3		1
Lumber	3	1	1	2	2	3	3	1	1	3	3	3	2	2	1	3	3
Pulpwood	1	1	1		1		3	2	2	1	3		1	1	3	2	2	2
Green manure	3	3	2	3	2	1	1	3	2	3	1	1	1	1	2	3	3	1
Craftwood	3	2		2	3	1	3	3	1	1	2		2		1	1	3	3
Food	3	3	3	3	3	3	3	3	3	2	2	3	1	3	2	2	3	1
Tolerance of:
Acid soils	1	1			2?		2?	1				3	3					3
Cold soil	2	2	1	3	2	1	2	2	1	3	3	2	3	1	3	3	2	3
Drought	3	3	2	3	3	2	2	1	1	1	2	1	1		1	1	2	2
Min. rain (mm)	120	150	100	100	150	60	100	30	50	75	150	60	60		25	60	75	100
Coppicing ability	2	1	1		1	1	1	2	1		1	1	1		2			1

Notes:

a	Acacia auriculiformis
b	Acacia mangium
c	Acacia mearnsii
d	Acrocarpus flaxinifolius
e	Albizia falcataria
f	Albizia lebbek
g	Calliandra callothyrsus
h	Casuarina equisetifolia
i	Dalbergia sissoo
j	Enterolobium cyclocarpum
k	Gliricidia sepium
l	Leucaena diversifolia
m	Leucaena leucocephala
n	Mimosa scabrella
o	Prosopis pallida
p	Samanea saman
r	Schizolobium parahyba
s	Sesbania grandiflora

Choice of fuelwood for home use involves many considerations. Local preferences dictate a wide array of species in the arsenal of the agroforester. Most simple stoves are designed to accomodate long pieces of wood that are fed into the stove as they burn. Most labor- and energy-efficient stoves are closed in order to minimize air intake, and so require specific, cut lengths. Split wood dries rapidly and is often favored over round wood, although marketing is conventionally by volume; not by weight. Irregular, heavily knotted woods (e.g.. many acacias, prosopis) are difficult to prepare or split as fuelwood, hut may he preferred for charcoal. Smokiness, ash content, explosive inclusions. thorniness, odor, and uniformity of burn can influence home fuelwood value. Many of these trails could be addressed profitably by the plant breeder and silviculturist. As an example. thornless mutants are found in several of the thorny mimosoids (Felker, 1979).

Green manure and nurse trees

Leguminous shrubs and trees are of increasing interest as sources of "green gold" (Curran, 1976) for the fertilization or nursing of both herbaceous and tree crops in the tropics. Green manuring of herbaceous crops is a sadly neglected area of tropical research. Legume trees like Leucaena end Sesbania can be continously coppiced for harvest of leaf meal. The clippings, which are high in N. can be placed directly around an interplanted crop, or "cut and carried" for incorporation prior to planting. Guevara, Whitney & Thompson (1978) showed that annual N yields of 0.5 t/ha can be obtained from Leucaena harvested every three months. Similar estimates may be inferred from earlier studies in the authors' laboratory. The availability of inorganic fertilizers has discouraged research on green manures in the tropics until recently. Definitive, quantitative data on N recovery and utilization from leguminous forage remains a serious need. Initial studies of R.A. Bradfield (personal communication) on leucaena green manuring of maize at IRRI were very promising. Guevara (1976) later quantified this relationship in Hawaii, recording excellent maize yields and effective recovery of about 46% of the N applied as leaf meal. An extensive demonstration of these methods is underway by the Philippine National Food and Agriculture Council. Legume trees of special merit for green manure research include the widely used Sesbania spp., Leucaena leucocephala and Gliricidia sepium (annually deciduous); also Acacia mearnsii, Albizia spp., Calliandra calothyrsus, and Mimosa scabrella.

TABLE 4: Properties of N₂-fixing trees in University of Hawaii international network trials.

Property	Genus and Species
	a	b	c	d	e	f	g	h	i	j	k	l	m	n	o	p	r	s
Specific gravity	.68	.65e1	.65	.63	.33	.58	.65	1.00	.68	.50	.75e	.55e	.54		.80	.52	.32	.42
Wood yield m³/ha.yr.	15	30	20e		40	5	50	15		8e	8e	25e	45		8e	15e	20	22
Average caloric value (Kcal/g)	4.8					5.2	4.6	5.0	5.0		4.9		4.6
Average annual height growth (m)	2.6	2.5	4e	2.0	5.0	1.4	6.0	2.1	2e	2e	25e	4.0e	4.5	4.5	2.5e	2.5e	1.9	3.3
Height at maturity (m)	30	30	25	60	45	30	10	30	30	30	10	20	20	15	20	45	30	10
DBH at maturity (cm)	60	25	50e	300	200	200	20	30	200	200	20e	20e	35	30	60	180	70	30

1e = estimated values.

Notes:

a	Acacia auriculiformis
b	Acacia mangium
c	Acacia mearnsii
d	Acrocarpus flaxinifolius
e	Albizia falcataria
f	Albizia lebbek
g	Calliandra callothyrsus
h	Casuarina equisetifolia
i	Dalbergia sissoo
j	Enterolobium cyclocarpum
k	Gliricidia sepium
l	Leucaena diversifolia
m	Leucaena leucocephala
n	Mimosa scabrella
o	Prosopis pallida
p	Samanea saman
r	Schizolobium parahyba
s	Sesbania grandiflora
a	Acacia auriculiformis
b	Acacia magnium
c	Acacia mearnsii
d	Acrocarpus flaxinifolius
e	Albizia falcataria
f	Albizia lebbek
g	Calliandra callothyrsus
h	Casuarina equisetifolia
i	Dalbergia sisoo
j	Enterolobium cycocarpus
k	Gliricidia sepium
l	Leucaena diversifolia
m	Leucaena leucocephala
n	Mimosa scabrella
o	Prosopis pallida
p	Samanea saman
r	Schizolobium parahyba
s	Sesbania grandiflora

The inter-planting of leguminous trees as nurse crop to other trees evolved out of the tradition of shading crops like coffee and cacao. Shade may in fact be a disadvantage offset by the N-rich leaf drop in many plantations. Among the major nurse legumes for plantation crops are Albizia carbonaria, Erythrina spp., Flemingia congesta, Inga spp. and Leucaena spp. (diversifolia, leucocephala and pulverulenta). Flemingia is notable for its tolerance of acid rubber plantation soils, as is Acacia auriculiformis.

Tree legumes can also he used as living fences or support systems for other crops. Studies. at the International Institute of Tropical Agriculture (IITA) (1979) have demonstrated the practicality of using Leucaena as living support for yams. winged beans and other crops (e.g., pepper. betel, vanilla, and passion fruit).

Forage

The leguminous trees commonly used for forage, following continuous clipping, include Cassia sturtii, Desmanthus virgatus, Leucaena leucocephala, and Sesbania grandiflora. Foliage of other species is palatable to animals and could he recovered during wood harvest, e.g., Acacia mearnsii, Albizia lebbek, Gliricidia sepium and Mimosa scabrella (see Table 4). Leucaena, the most intensively studied of the species listed above, can produce 10-15 tons (dry matter-) of forage per hectare annually (Brewbaker et al., 1972) when harvested regularly. The value of the foliage as co-product in fuelwood or pulpwood her vest may be great enough in the case of Leucaena to encourage use of chip-vacuum, leaf-meal recovery machines.

Many of the 600 Acacia spp. hear phyllodes (expanded petioles) as mature leaves that are generally fibrous and unpalatable, Mimosine (in all Leucaena spp.) and other alkaloids occur in some tree legumes and require caution in their use as forage. Breeding and management of the forage (e.g.. silage preparation) may offer solutions to these problems (González, Brewbaker & Hamill, 1968; Rosas, Quintero & Gómez, 1980).

University of Hawaii trial network for N2-fixing trees

The US National Academy of Science reports on Leucaena and on tropical legumes prompted an expansion of genetic research in Hawaii on N₂-fixing trees, previously confined to Leucaena and Acacia koa. A major thrust of the expanded studies is to determine relative biomass yields of different species and varieties of leguminous trees. A trial network for Leucaena was initiated in 1978, and expanded with USDA support in 19X0 to include other species.

The major species chosen for our studies are summarized in Tables 3 and 4. All are considered relatively fast growing, with most species exceeding 15 m³/ha per yr of wood. Most are hardwoods with high intrinsic value as fuel or pulpwood, and several are valued for forage or lumber and craftwood. Acid and unusually arid soils, along with waterlogged and saline soils, present primary challenges to the forester. In this study Acacia auriculiformis was chosen for relative tolerance to acidity and Prosopis pallida for relative tolerance to aridity.

Yield trials are planted with dense spacing (5000 or 10,000/ha) using 3-to 4-month-old seedlings transplanted into small plots (minimally 28 m²). Trials use the augmented block design (Federer & Raghavarao, 1975), and include several replications of 10-15 species, but can include additional unreplicated plots of other species or treatments. This is a flexible design that accommodates diverse entries and treatments at different locations, yet permits the pooling of replicated data for calculations of variety x location and error terms.

Initial results of such international trials Leucaena have been gratifying. Giant varieties Leucaena provide some of the fastest growth and greatest versatility of the tree legumes, probably equal to any nonlegume.

Research imperatives

With perhaps a thousand potentially significant N₂-fixing trees to study in the tropics, where should research emphasis be placed? It seems wise to focus on species providing both forage and fuelwood to the small farmer. Few nonlegumes bear consideration, and species achieving less than 2 m annual growth should be excluded. Thorniness must be considered undesirable, despite the protection it gives against animal depradation. The following dual-purpose species appear to deserve extensive collection, genetic evaluation, and site adaptability studies:

Acacia spp. (see Table 2)
Calliandra calothyrsus
Gliricidia sepium
Leucaena leucocephala
Prosopis spp. (Table 2)
Sesbania grandiflora

References

Barney, G. (1978) The nature of the deforestation problem: trends and policy implications. In: Proceedings, US strategy conference on deforestation. US State Dept. Washington, D.C., USA.

Brewbaker, J.I., Plucknett, D.L. & Gonzales, V. (1972) Hawaii Agric. Exp. Bull. 166. 29 pp.

Brewbaker, J.L., (Ed.) (1980) Giant Leucaena (koa haole) energy tree farms: An economic feasibility analysis for Island of Molokai. Hawaii Natural Energy Institute, Publ. No. 80-160. 90. Pp.

Brewbaker, J.L., & Hutton, E.M. (1979) Leucaena versatile tropical tree legume. In: New agricultural crops, G.A. Ritchie (Ed.). Assn. Adv. Sci., Westview Press. Boulder. CO. USA. Pp. 207-257.

Curran, H., (1976) Giant ipil-ipil: Green gold for the tropics. US Peace Corps, mimeographed.

Federer, W.T. & Raghavarao, D. (1975) Biometrics 31. 29-35.

Felker, P. (1979) Mesquite, an all-purpose leguminous arid land tree. In: New agricultural crops, G.S. Ritchie (Ed.), Amer. Assoc. Adv. Sci., Westview Press, Boulder CO. USA. Pp. 207-257.

Food and Agriculture Organization (FAO) (1980) Forestry for rural communities. Forest Dept. Publ., FAO, Rome, Italy, 53 pp.

González, V., Brewbaker, J.I., & Hamill, D.E. (1967) Crop Sci. 7, 140-143.

Guevara, A.B. (1976) Management of Leucaena leucocephala (Lam.) de Wit for maximum yields and nitrogen contribution to intercropped corn. PhD thesis. Univ. of Hawaii, Honolulu. Hawaii, USA.

Guevara, A.B., Whitney A.S. & Thompson, J.R. (1978) Agron. J. 70. 1033-1037.

International Institute for Tropical Agriculture (IITA) (1979) Annual report. IITA. Ibadan, Nigeria.

National Academy of Science (NAS) (1977) Leucaena: promising forage and tree crop for the tropics. NAS. Washington, D.C., USA, 115 pp.

National Academy of Science (NAS) (1979) Tropical legumes: Resources for the future, NAS, Washington, D.C., USA, 115 pp.

National Academy of Science (NAS) (1980) Firewood crops: Shrub and tree species for energy production. NAS, Washington, D.C., USA 237 pp.

Revelle, R. (1980) Science 209, 164-174.

Rosas, H., Quintero, S.O. & Gómez, J. (1980) Leucaena Research Reports 1,17 pp.

Stewart, W.D.P. (1967) Science 158, 1426-1432.

Woodwell. G. (1978) Carbon dioxide-deforestation relationships. In: Proceedings, US Strategy Conference on Tropical Deforestation, US State Dept., Washington. D.C., USA.

World Bank (1978) Forestry sector policy paper. World Bank, Washington, D.C., USA, 65 pp.