Overstory #120 - Seed Collection
A high degree of diversity is common in the Tropics, especially in the humid lowlands. In single-species operations, such as the extensive plantations of Pinus spp., natural diversity is irrelevant. But organizations that service a wide variety of users with diverse needs, and ecosystem restoration projects where nature dictates the species to plant, cannot ignore diversity. Poorly known species and species with difficult reproductive characteristics must be accommodated. With a good understanding of biology and ecology, innovative thinking, and a little luck, most of these new species can be collected and grown. In some cases research can provide answers, but for a few species, and for a variety of reasons, consistent collection and processing of viable seeds may be impossible. In a 14-year study of species suitable for revegetation of bauxite mine lands in Trombetas, Brazil, 600 species were evaluated and 160 were grown and outplanted; only 89 taxa demonstrated acceptable survival and growth during the first two years (Knowles and Parrotta 1995).
If diversity is a distinguishing feature of tropical forests, diversity also characterizes seed collection methods in the Tropics. Because the number of species collected is large and the objectives and budgets of seed collecting organizations are varied, a wide array of collection methods must be used. Redundancy is an important feature of high biodiversity; often several alternative species are available to fill any particular need. Seed collectors may not always be successful in harvesting seeds of all the species desired, but by using all the tools available, by taking advantage of seed opportunities as they arise, and by substituting species with similar properties, they can consistently offer an adequate range of seeds and seedlings to users.
This article reviews some of the challenges to seed collection and processing in the Tropics and the approaches to meeting these challenges.
One of the most important differences between tropical and temperate forests is the high degree of species diversity in tropical forests. While a tract of temperate forest might contain 25 tree species, a tropical forest of similar size might easily support 10 times that number. Indeed, there are probably more than 50,000 arborescent species in the Tropics. As byproducts of this vast biological diversity and a broad range of climate, economic conditions, and political systems, tropical tree nurseries and seed-collecting organizations vary enormously in structure and in methods employed to collect and process seeds. These nurseries and organizations range from the most sophisticated and mechanized operations to tiny nurseries that use only hand labor and locally manufactured materials. All use valid and necessary means to solve the unique problems of each local situation.
Seed suppliers that produce seed for sale both within their own countries and for export are required to produce the highest quality product possible. It must contain little debris and be free of insects and weed seeds. Superior genotypes bring premium prices, but all lots must be forest-run (representing the average for wild trees) or collected from trees of selected phenotypes. There is constant pressure to offer as many species as possible, but costs tend to escalate with the number of species maintained in collections. These operations may employ one to many people, have varying degrees of mechanization, and be government affiliated or private.
The most common objective of seed collection in the Tropics is to support local tree nurseries. These nurseries are the principal source of tree seedlings for homeowners, cities, farmers, and conservation and forest management organizations who use the seedlings for ornamental, agricultural, agroforestry, conservation, and silvicultural purposes. They typically maintain a relatively large inventory of the few species that are in continual demand (all of which are reliable seed producers) and a group of other species that changes as seed collection opportunities present themselves. Seeds are usually collected by nursery employees or purchased from local seasonal collectors. Seeds are rarely stored for more than 1 year. Seed quality and handling are highly variable. Because wages are low in most of the Tropics and in some places the objective of providing employment supersedes that of producing seedlings, labor is often substituted for equipment or improvements requiring capital investment. Unfortunately, product quality and efficiency may decrease.
It is important to remember that fruit characteristics evolved to facilitate a seed dispersal strategy. Often a seed collection, extraction, or germination technique has simply mimicked the natural process.
From the Ground
The mainstay of seed collection is and will remain the picking up of the fruits or their seeds from the ground after they fall. This method is especially convenient for species with large or conspicuous fruits or seeds such as Melia azedarach L., Ormosia krugii Urban, and Terminalia catappa L. Protective pods enable collection of fruits of such species as Crescentia cajute L., Hymenaea courbaril L., Pterocarpus macrocarpus Kurz, and Senna spectabilis (DC.) Irwin and Barneby, weeks to months after fruit-fall. Even the seeds or fruits of relatively small-seeded species, such as Bucida buceras L. and Petitia domingensis Jacq.., can be collected from the ground if it is bare or paved, or if a tarp is placed under the tree just before fruit-fall. In those circumstances, vacuuming or sweeping can sometimes speed the collection process.
From the Tree
Another common means of collecting fruits and seeds is picking them from the trees. It is faster in many cases than collecting from the ground and it keeps the seeds cleaner. In addition, the seeds of many species are too small to pick up from the ground, and others are dispersed widely or consumed by animals and insects before they fall. Short trees may allow hand picking. Production can be accelerated by flailing into a basket attached to a picker's waist.
Picking must occur between physiological ripening (timed so the fruits will ripen in storage) and release by the tree or consumption by predators or dispersers. Ripeness is usually indicated by a color change of the fruit from some shade of green to an indicator color such as brown or red. In some cases, animal predation indicates ripeness. The sight of cockatoos feeding on the seeds indicates that Agathis seeds are ripe (Whitmore 1977). The seeds of many species, such as Albizia lebbeck (Kunth) Harms and Melia azedarach L., remain in their fruits on the trees for weeks or months, greatly facilitating collection. Many of the eucalypts carry large quantities of fruits for extended periods that open quickly after fires (Cremer and others 1978). These fruits will also open after the clipped twigs dry. Species with serrotinous cones, such as Pinus patula Schiede & Deppe, produce cones that remain on the tree with viable seeds 1, 2, or more years after ripening (Wormald 1975).
Most species are too tall for hand picking. Pole pruners provide a convenient and inexpensive way to collect fruits from 2 m to about 9 m above the ground. Stepladders and easily portable, straight or extension ladders up to 7 m are also used. Fruits or seeds that shatter or detach easily, or seeds that are ejected from their fruits, can be collected by placing a tarp under the tree and flailing the tree with a long pole. In similar approaches, trees are shaken by hand (if small), by mechanical shaker, or by attaching a chain or rope to the trunk as high as convenient and to a vehicle or another tree, then jumping up and down on the chain or rope or pulling back and forth with a vehicle. A particular benefit of collecting seeds by shaking is that for species such as Cordia alliodora (Ruiz & Pavon) Oken, in which all the fruits or seeds do not mature at the same time, shaking releases the mature fruits while the immature fruits remain on the tree (Greaves and McCarter 1990).
When short-statured species, such as Acacia farnesiana (L.) Willd., Hibiscus tiliaceus L., and Moringa oleifera Lam., and precociously flowering species such as Spathodea campanulata Beauv., invade cleared areas, they are short enough to pick by hand or pole pruner. Trees planted in open-grown situations sometimes grow low enough for easy harvest. Open-grown Swietenia spp. will usually bear many of its fruits within 9 m of the ground but does not do so in closed forests. Many species will continue to bear fruits after pruning to a low crown. Fruit trees, including Mangifera indica L. and Citrus spp., are routinely managed this way. Crown-shaping trials with many tropical forest species should be conducted.
The seeds of timber trees in natural forests are often borne near the tops of very tall trees. Traditionally, these seeds are collected from trees felled by logging operations. The method is excellent, if available. However, felling can shatter seed clusters and produce enough slash to hinder fruit retrieval. In some areas, trees have been destructively felled to obtain the seeds (Britwum 1973). Generally considered unacceptable ecologically and economically, this method may be beneficial if the seed trees are scheduled for removal to improve the stand.
The challenge of collecting seed from tall trees has inspired a number of approaches. Arrows have been used to draw lines over limbs to cut down or shake down small quantities of seed. Traditionally, local climbers were hired to ascend the trees and pick the desired seed in quantity with little or no safety equipment. Any conscientious collector would now require the use of safety harnesses and belts, safety lines, and helmets. Tree bicycles, tethered sectional ladders, and other climbing aids can greatly accelerate an ascent up the tree bole. Once up the tree, a collector often uses a pruning pole to clip fruits on the ends of limbs or to saw off fruit-bearing limbs. Finally seed collectors with generous budgets use high-lift buckets on hard and nearly level ground to collect fruits from tall trees.
Widely Scattered Seed Trees
In tropical moist forests, seed trees of many species grow more than 1 km apart. Seed-producing adults of heavily exploited and rare species can often be very difficult to find. Random searches can be incredibly time consuming. Traditionally, seed collectors maintain a mental inventory of known seed trees or potential seed trees they have sighted during excursions through the forest over the years. Although it takes many years to accumulate this degree of experience, the system works well as long as these inventories are handed down to succeeding generations of collectors. The alternative is to maintain a written inventory of potential seed trees of at least the critical species along with maps of their locations. Modern technology can add efficiency and accuracy to this time-consuming process through the use of computer data bases and global positioning satellite (GPS) technology.
A number of species hybridize freely with members of the same genus or varieties within the same species to produce undesirable or unpredictable offspring. The Swietenias have hybridized freely in Puerto Rico and the seeds of Eucalyptus robusta Sm., obtained from Brazil many years ago, introduced many hybrids that were generally inferior to the pure species. The timber tree Hibiscus elatus Sw. has been reported to hybridize with the shrubby Hibiscus pernambucensis Arruda in Jamaica (Adams 1971). In the case of Swietenia, the hybrid seedlings can usually be separated from the parent species by leaf size. With B. robusta, sorting is impossible. The problem is solved by collecting from well-identified and isolated seed trees.
Unknown or Unpredictable Fruiting Seasons
In that portion of the Tropics with a strong wet-dry seasonal cycle, almost all species flower and fruit in certain well-defined seasons. Phenological studies can document these seasons and seed collection can be planned accordingly. However, in many species the flowering and fruiting dates may vary somewhat (Greaves 1978) and the level of fruiting may vary tremendously from year to year, depending on seasonal rainfall amounts and patterns, and other factors such as wind or insect damage. Seed years are theoretically predictable from climatic conditions and can be successfully predicted from flowering and fruit set for a few species, such as Pinus caribaea sensu Small, non Morelet, which flowers one year and bears seeds the next.
In those portions of the Tropics with a relatively even distribution of rainfall, many species, such as Roystonea spp., Ficus citrifolia P. Miller, and Hibiscus elatus Sw., flower and fruit irregularly throughout the year. Some species, such as Leucaena leucocephala (Lam.) de Wit, bear seed more or less continuously in moist habitats, but seasonally in habitats with a strong wet-dry cycle. Vochysia hondurensis Sprague in Costa Rica bears fruit twice per year (Nichols and Gonzalez 1992a, 1992b). Fruiting in some species is thoroughly unpredictable; individual trees fruit irregularly from year to year and by seasons and are not synchronized with others of their species. However, this can be an advantage. If seeds are unavailable on one tree, they may be present on the next tree; or if seeds are unavailable in one locality, they may be available a few kilometers away. Byrsonima spicata (Cay.) Kunth, Cordia sulcata DC., and Buchenavia tetraphylla (Aubl.) Howard illustrate this behavior. Although Swietenia macrophylla G. King is usually seasonally synchronized even outside its native range, a few unsynchronized individuals produce some seeds during most of the year. Continuous flowering and fruiting and discontinuous/irregular year-round fruiting is a natural strategy to avoid overloading the demand for pollinators and seed dispersers. Moreover, having a small percentage of the population out of synchrony helps avoid loss of regeneration to irregularities in normal rainfall patterns. To solve the problem of collecting from species with this diverse behavior, phenology should be recorded and collection activities should be planned by species. Collectors should not become discouraged by a few failures.
Delayed and Rare Fruiting
Many species, such as Bertholletia excelsa Humb. & Bonpl., do not bear fruit until they become large canopy dominants, a process that may take 50 years or more. In plantations, the process can be shortened to 15 to 25 years and grafted stock will bear fruits in as little as 6 years (Ferraz 1991). Seed orchards exploit the tendency for open-grown trees to bear fruit more quickly and prolifically than forest trees. The abundance and ease of collection from open-grown trees has often led to excessive collection from phenotypically unproven trees in pastures and along streets. Lagerstroemia specious (L.) Pres. begin fruiting in as little as 3 years, but do not bear viable seeds until about 15 years (Food and Agriculture Organization 1957). Most species of Bambusa flower and fruit in regional synchrony only once every few decades, and Corypha umbraculifera L., a palm, flowers once at the end of its long life and then dies.
The advantages to collecting seeds of species that can be stored during bumper crop years are numerous. The cost per unit of seed is lower; fewer seeds are damaged by insects; and the seed usually germinate at higher rates (Lamb 1993). A number of healthy exotic trees, such as Araucaria heterophylla (Salish.) Franco in Puerto Rico, do not produce seeds in their new habitats (Francis 1987), and seeds must be imported each season. Ringing, shallow girdling, stem strangulation, stem bending, root pruning, and water supply restriction has shown some promise in promoting seed production, although these methods ultimately injure the trees (Rudolf and others 1974).
Rodents, monkeys, birds, bats, and grazing animals can quickly eliminate a seed crop in a limited area. Parrots in Central America can consume an entire seed crop of Acacia aneura F. Muell. before it ripens (Willan 1995). Although fences, screens, scarecrows, reflectors, and noise makers can successfully reduce or eliminate seed predation, they are usually practical only for seed orchards or concentrations of seed trees. Sometimes seeds can be harvested after they have become viable but before they or the fruits become attractive or accessible to animals. When forest species are scattered, collecting more widely and intensively appears to be the only way to obtain the needed seed stocks. In temperate areas, squirrel caches can be robbed; and in the Tropics, seeds are separated from the manure of predators that have been feeding on the fruit of the desired species.
Most species are attacked to some degree by seed insects. Occasionally, species such as Prosopis juliflora (Sw.) DC. and Triplochiton scleroxylon K. Schum. are so seriously attacked by insects that propagation is limited (Brookman-Amissah 1973, Marrero 1949). In some cases, insecticides can be used to prevent attack and assure good seed crops. Zanthoxylum flavum Vahl seeds in Puerto Rico are reduced to less than 5 percent viability by a seed weevil (Francis, personal communication 1994; Marrero 1949). A conservation organization was able to produce seeds free of insects and with good germination by spraying with insecticide (Rivera, personal communication). Many types of seeds should also be treated by fumigation, cold treatment, or insecticide application to eliminate insect damage during drying and storage.
Short Period of Availability
For various reasons, fruits or seeds of many species are ripe and available on trees for a very short time, often just a few days. The fruits of Hyeronima oblonga Muell. Arg. and H. alchornioides Allem. Diss. fall 3 to 4 days after maturity (Nichols and Gonzalez 1992a, 1992b). In the final stage of ripening, Pinus caribaea sensu Small, non Morelet cones change color from green to brown, the cones open, and the seeds are dispersed quickly (Greaves 1978). Frequent field checks are essential to best time seed collection. Because individual trees of a species are often not closely synchronized, collecting from tree to tree can extend the collection season. Often, the collector can lengthen the period of seed collection by moving up an elevational gradient or across a moisture gradient. Some species, such as Maesopsis eminii Engl. and Pouteria spp., picked up just before ripening, ripen in storage; thus the picking season is lengthened by a few days. The collector must know the species' traits, because some species, such as Cordia alliodora (Ruiz & Pavon) Oken, stop ripening as soon as they are detached from the tree (Greaves and McCarter 1990). The collector should also study how long seeds of critical species will remain viable on the ground after fruitfall. Collecting diptocarp seeds from the ground, for example, must be carefully timed because a delay of a few days can result in loss of viability (Domingo 1973).
Collection and processing problems are species-driven. Hence, the greater the number of species handled, the greater the number of problems that must be solved. Unfortunately, some species are planted because their seed collection and management are easy, not because they are the best species available to meet the need. Many superb species are rarely or never planted because their seeds are difficult to collect or use. As experience in seed collection and management increases, many exciting species will be added to plantation inventories. Increasing the number of species we are able to successfully reproduce by seed will facilitate the difficult task of rehabilitating damaged ecosystems.
Adams, C.D. 1971. The blue mahoe and other bush: an introduction to the plant life of Jamaica. Singapore: McGraw-Hill Eastern Publishers (S) Ltd. 159 p.
Britwum, N.V.L. 1982. Seed problems of indigenous plantation species in Ghana. In: Seed problems. Paper 7. Stockholm, Sweden: International Union of Forest Research Organizations, Working Party S2.01.06. [not paged]. Vol 2.
Brookman-Amissah, J. 1973. Seed problems as they affect foresty practice in Ghana. In: Seed problems. Paper 32. Stockholm, Sweden: International Union of Forest Research Organizations. Working Party S2.01.06. [Not paged]. Vol 2.
Cremer, K.W., Cormer, R.N., Florence, R.G. 1978. Stand Establishment. In: Willis, W.E.; Brown, A.G., eds. Eucalypts for wood production. Adelaide, Australia: Commonwealth Scientific and Industrial Research Organization: 81-135.
Domingo, I.L. 1973. Seed problems in the regeneration of the Philippine diptocarp forests. In: Seed problems. Paper 32. Stockholm, Sweden: International Union of Forest Research Organizations. Working Party S2.01.06. [Not paged]. Vol 2.
Food and Agriculture Organization. 1957. Tree planting procedure in tropical Asia. Forestry Development Paper 11. Rome, Italy: Food and Agriculture Organization of the United Nations. 172 p.
Francis, J.K. 1987. Araucaria heterophylla Salisb. Franco Norfolk-Island-pine. Res. Note SO-ITF-SM-11. New Orleans: US Department of Agriculture, Forest Service, Southern Forest Experiment Station. 4 p.
Greaves, A. 1978. Description of seed sources and collections for provenances of Pinus caribaea. Tropical Forestry Paper 12. Oxford, UK: University of Oxford, Commonwealth Forestry Institute. 98 p.
Greaves, A.; McCarter, P.S. 1990. Cordia alliodora: a promising tree for tropical agroforestry. Tropical Forestry Paper 22. Oxford, UK: University of Oxford, Commonwealth Forestry Institute. 37 p.
Knowles, O.H.; Parrotta, J.A. 1997. Phenological observations and tree seed characteristics in an equatorial moist forest at Trombetas, Para State, Brazil. In: Lieth, H.; Schwartz, M.D., eds. Phenology in seasonal climates I. Leiden, The Netherlands: Backhuys Publishers: 67-86.
Lamb, A.F.A. 1993. Pinus caribaea. Fast growing timber trees of the lowland tropics 6. Oxford, UK: Oxford University, Department of Forestry, Commonwealth Forestry Institute. 254 p.
Marrero, J. 1949. Tree seed data for Puerto Rico. Carribean Forester. 10: 11-30.
Nichols, D.; Gonzales, E. 1992a. Especies nativas y exoticas para la reforestacion en la zona sur de Costa Rica. San Jose, Costa Rica: Organizacion Par Estudios Tropicales and Direccion Tropical. 84 p.
Nichols, D.; Gonzales, E. 1992b. Especies nativas y exoticas para la reforestacion en la zona sur de Costa Rica. Memoria del II Encuentro sobre Especies Forestales. San Jose, Costa Rica: Cedicion Universidad Estatal a Distancia, Organizacion Par Estudios Tropicales and Direccion General Tropical. 73 p.
Rudolf, P.O., Dorman, K.W., Hitt, R.G., Plummer, A.P. 1974. Production of genetically improved seeds. In: Schopmeyer, C.S., ed. Seeds of woody plants in the United States. Agric. Handb. 450. Washington, DC: US Department of Agriculture: 53-74.
Whitmore, T.C. 1977. A first look at Agathis. Tropical Forestry Papers 11. Oxford, UK: University of Oxford, Commonwealth Forestry Institute. 54 p.
Willan, R.L. 1995. Problemas fitosanitarios en el abasticimento de semillas. Serie Materiales de Ensenanza 32. In: Jara, L.F., ed. Programas de abastecimiento de semillas forestales. Turrialba, Costa Rica: Danida Forest Seed Centre and Centro Agronomico Tropical de Investifacion y Ensenanza. [not paged].
Wormald, T.J. 1975. Pinus patula. Tropical Forestry Papers 7. Oxford, UK: University of Oxford, Commonwealth Forestry Institute. 212 p.
This article was adapted with the kind permission of the author and publisher from:
Vozzo, J.A. (Ed). 2002. Tropical Tree Seed Manual. USDA Agriculture Handbook 721.
About the author
John K. Francis grew up on a small dairy farm in Wendell, Idaho. Perhaps because this was desert country with few trees, he always wanted to plant and manage trees and a career in forestry followed naturally. He received a BS in Forest Management from the College of Forestry at the University of Idaho where he also received a PhD in 1974, specializing in forest soils. For most of his career he has worked as a research forester in tropical silviculture at the USDA International Institute of Tropical Forestry, Rio Piedras, Puerto Rico (now stationed at the Shrub Sciences Laboratory in Provo, Utah). His interests include growth, yield, management, and site adaptability of upland and bottomland hardwoods and conifers; soil nutrient depletion by intensive harvest; silvics of timber and nontimber trees; and thamnics of tropical shrubs.
His publications include Silvics of Native and Exotic Trees of Puerto Rico and the Caribbean Islands (USDA Forest Service, 2000). He is currently editing a national shrub manual for the United States. He can be contacted at: RMRS Shrub Lab, 735 North 500 East, Provo, UT 84606; Tel: 801-356-5137; Fax: 801-375-6968; E-mail: firstname.lastname@example.org.
Related editions to The Overstory