Summary

The first oil palm seeds were brought to Tropical America from Malaysia, Indonesia and Sierra Leone by the United Fruit Company between 1926 and 1929. Seeds of some of these materials were distributed for the first commercial plantings in Central and South America for the early developments. Starting in 1967, through exchange programs, genetic materials were introduced by ASD to Costa Rica for further improvement and selection, from renowned oil palm breeding research stations such as Chemara, SOCFIN, MARDI, Banting, Dami, La Mé (IRHO), Lobe (Cameroon) and Kade (Ghana). The ASD germplasm pool has also been enriched with collections from Tanzania, Cameroon (Bamenda), Mobai, Sierra Leone, Uganda, Zambia and Malawi. The ASD breeding program's efforts have been concentrated mostly on improving advanced Deli duras introduced from the above five Asian breeding research stations, and the AVROS, Ekona and Calabar pisifera sources. As a result of this, significant progress has been achieved mainly with the Deli duras from Chemara and Dami, and with the Ekona and Calabar pisifera sources. ASD Deli x AVROS planting material was first used in commercial plantations in Costa Rica in 1976, whereas Deli x Ekona and Deli x Ghana (Calabar) seeds were initially sold in 1989. To date there are 181,000 hectares planted with ASD seeds in Latin America and 245,000 hectares in Southeast Asia and India. The fact that some experimental plots and small commercial plantations growing under very suitable environmental and management conditions have produced up to 40 tons per hectare in peak years, shows the high yielding potential of ASD planting materials. However, because most plantations in Latin America are located in regions with important climatic and soil limitations for oil palm cultivation, average commercial yields range from 20 to 30 tons of FFB per hectare per year. Based on carefully-followed pollination techniques, ASD guarantees 99.9% purity of teneras. ASD's current production capacity is around 35 million seeds per year.

Introduction

ASD de Costa Rica started the commercial and technological development of the oil palm on the American continent since the late 1920s, with the introduction of the first oil palm seeds of known origin. For about 50 years now, ASD's goal has been the development of the industry through continuous research. Under this scheme, ASD has gathered a broad germplasm collection of both Elaeis guineensis and E. oleifera. The results of these efforts and a strong commitment to achieve the highest quality standards that guarantee over 99.9% teneras, have made ASD a very reliable and successful international oil palm seed supplier, playing a key roll in the industry expansion worldwide.

The history of germplasm introductions to Costa Rica

The first documented introduction of oil palm of known origin into the Americas goes back to 1926, when the Panama Division of the United Fruit Company (former ASD's parent company) brought the first seeds from Malaysia. In 1927 the Lancetilla Experimental Station in Honduras also imported seeds from Asia and Africa. Following these introductions, in 1929 the Guatemala division imported 1,000 seeds from Sierra Leone (Richardson 1995).

Starting in 1936, open-pollinated seeds from Lancetilla, Honduras were distributed to the first commercial plantings in Cuba, Guatemala, Costa Rica, Colombia, Ecuador and Peru. Despite these early developments, the oil palm industry in the Americas did not grow as much as in Asia.

Richardson (1995) describes in detail how ASD's genetic resources were introduced into Costa Rica for further selection and improvement. Most of these introductions were obtained through the exchange of E. oleifera from Central America for E. guineensis from the most important oil palm breeding stations in Africa and Southeast Asia. In December 1967 the following materials were introduced from Chemara Research Station Malaysia in cooperation with B. Gray, A. H. Green and J. J. Hardon:

• 6 crosses Deli x URT
• 6 crosses Deli x BM 119
• 2 crosses BM 119 x BM 119
• 1 cross URT x URT
• 3 crosses Chemara Delis
• 3 crosses H&C Delis

Exchange programs with Unilever (Cameroon) began in July 1967, when nine seed lots of wild palms from the Bamenda Highlands were shipped to Central America. In May 1969 G. Blaak sent 14 crosses, mainly of Ekona materials. These crosses included outstanding pisifera parents in the accession CAM 236 (2/2311T x 3AR/7239T).

The SOCFIN (Societé Financière de Caoutchomes) Research Department sent to Costa Rica pollen of Sibiti pisiferas, Deli dura seeds, Yangambi, and La Mé T x T crosses in October, 1970. In 1975 an extensive program of crosses of Central American' Oleiferas with pollen from the former IRHO(Institut de Recherches pour les Huiles et Oléagineux) was started. As a result of this cooperative project, five accessions each of La Mé, Yangambi, and Nigerian origin TxT and TxP crosses were delivered to the United Fruit breeding project in 1979.

Similarly, seed exchange programs were arranged with: i) S. C. Ooi of MARDI(Malaysia), in 1977; ii)Tam Tai Kin of Banting (Malaysia), in 1977; iii) E. A. Rosenquist of Dami(Papua New Guinea), in 1977 and iv) J. B. Wonkyi-Appiah of Kade (Ghana), in 1977.

In 1978, G. Blaak prospected for oil palms in the Kigoma district of Tanzania. Some of these materials were characterized by "paper thin" shells, and most had good bunch composition. Five accessions of this origin were established at the ASD Coto Research Station. Seeds were also collected from palms in Mobai and Sierra Leone by G. Blaak in July, 1979. These palms were also established at the Coto research station. More recently, collections from the Entebbe Botanic Garden (Uganda), Zambia and Malawi have been added to ASD germplasm pool. The introductions made to Costa Rica since 1969 are summarized in Table 1 .

ASD'S Breeding populations

Deli Dura Mother Palms

The historical pathways of the different sources of Deli duras introduced to Costa Rica, denominated by Rosenquist (1985) as breeding populations of restricted origin (BPRO), are shown in Figure 1 .

The origin of the Costa Rican Delis goes back to the four original palms planted in the genetic garden in Bogor, Java in 1848. Seeds from these palms were planted along the roadside within the tobacco plantations in Deli, Indonesia. The seeds for the first commercial plantations in Sumatra, Malaysia and Honduras were originated from these ornamental palms.

Selected palms from commercial plantations formed the population base of Deli duras for further selection and genetic improvement in Sumatra (Bangun Bandar, SOCFIN), and Malaysia Harrison and Crosfield (H&C), Chemara and MARDI. H&C continued breeding Delis in Papua New Guinea until 1996 (Rosenquist 1985; Breure 1985). Advanced breeding materials from these five programs were introduced to Costa Rica following different paths as shown in Figure 1 .

From the original introductions of Deli dura progenies, considerable breeding progress has been achieved at the ASD Palm Research Program in Coto, Costa Rica. On Table 2, the characteristics of the germplasm originally received as compared with the following generations planted are presented. It is important to note how subsequent generations show much better bunch composition and yield potential than the original material.

Chemara and Dami dura materials apparently carried much more variability for further progress in Costa Rica, since the selection gains were superior in these materials than the progress achieved with the Banting and SOCFIN germplasm. It is clear that considerable improvement can still be achieved within the Deli dura population, despite its narrow genetic base, as pointed out by Rosenquist (1985).

AVROS: an outstanding source of pisiferas

Materials from two AVROS sources were introduced into Costa Rica as shown in Figure 2 . One set of TxT crosses came from the Banting (H&C) program in Malaysia, known as the BM 119 series, and the second was originated from the MARDI program.

The origin of the AVROS (Algemene Vereniging van Rubber-planters ter Oostkust van Sumatra) pisiferas goes back to the Eala Botanical Garden in Zaire, where the famous `Djongo' (the best) palm was selected. Progenies of the `Djongo' palm were introduced into Indonesia in 1922 for further improvement, and the famous SP540 palm was selected. Lines were developed from the SP540 pisifera, and are widely used for seed production in Malaysia and Indonesia.

The H&C AVROS pisiferas introduced to Costa Rica carried genes from another germplasm source from Western Africa, which had been introduced in Sumatra (Bangun Estate), Indonesia, in 1922. Crosses derived from several selection cycles carried out in the Polonia and Aek Pantjur plantations in Indonesia were imported into Malaysia in 1957, where the progeny BM119 was bred at the Banting Experimental Station. Crosses of this origin were selected by J. J. Hardon and introduced in Costa Rica as the BM 119 series in 1968 and 1980. Another set of BM119 pisiferas came from the MARDI program in 1979 ( Figure 2 ).

Phenotypic evaluation of tenera siblings in T x T progenies for the next generation of pisiferas is carried out systematically in ASD's breeding program. From the gains made by the selection resulting in the F1 generation of TxT progenies, the potential of pisiferas to transfer good characteristics to their progenies in the progeny tests is demonstrated. In Table 3, the evaluation of bunch characteristics of different ASD sources of pisiferas is presented. It is interesting to note that within the AVROS population little progress has been achieved in terms of O/B (gain 2.7%) in the selection cycles performed in Costa Rica. This is not surprising, since ASD AVROS progenies are very uniform, as a result of seven cycles of selfing,six carried out in Indonesia and Malaysia, and one in Costa Rica. Hence, further genetic progress of the AVROS germplasm is restricted.

Deli x AVROS is the classic material sold by ASD, and also the best-known and most-widely used worldwide. Under good management and agro-climatic conditions, it starts yielding 22 to 26 months after field planting, and FFB yields often surpass 30 tons/ha/year in the 6th year. It produces good-sized bunches, and, with proper management of harvest and mill operations, oil extraction rates of 24 to 25% are common.

Ekona and Calabar pisiferas as new alternatives

The Ekona population was developed at Unilever's Lobe Breeding Program, in Came-roon. Fourteen progenies provided by G. Blaak were introduced into Costa Rica in 1970. ( Figure 3 ). This population was much more variable than AVROS, since considerable progress was achieved (17.1% gain in O/B) after the selection cycles carried out in Costa Rica. One important characteristic of the Ekona population is its high oil to bunch content (Table 3).

Table 4 shows the results of progeny test of different pisifera sources in Coto, Costa Rica. The superiority of Ekona material is confirmed in terms of oil per hectare.

The Deli x Ekona materials were released for commercial planting by ASD in 1989. The Ekona have other attractive characteristics such as moderately short trunks, large bunches and early high yielding. Some Deli x Ekona crosses have shown tolerance to Fusarium Wilt in a test conducted in Brazil, and have performed well in areas with three to four months of moderately dry and cold weather.

The Calabar population was introduced to Costa Rica in 1977 with the cooperation of J.B. Wonkyi-Appiah of Kade, Ghana. These materials were originated from the old West African Institute for Oil Palm Research (WAIFOR) in Nigeria. The pedigree of the most important Calabar progeny is shown in Figure 4 . The single grandparent of GHA648 (32.3005T) performed quite well as male parent on Deli females in the Sabah breeding program in Malaysia (Rajanaidu et al. 1985 and Chan et al.1985), the Kade program in Ghana (Wonkyi-Appiah, 1974), and in the NIFOR program in Nigeria (Okwuagwu 1985; van der Vossen 1974).

The Calabar germplasm has considerable variability for further improvement (8.6% gain in O/B, Table 3), and in progeny tests has yielded slightly better than Deli x AVROS in terms of oil per hectare (Table 4). This material, which was released for commercial planting by ASD in 1989, has shorter trunk and leaves than Deli x AVROS, which makes it possible to reduce the plant spacing to only 8.5 x 8.5 m (160 plants per hectare). It also shows high early yields, but smaller bunches. Its bunch oil content is as good as that of Deli x AVROS. Besides having good production characteristics, this planting material has performed quite well in regions with a three to four month dry season and periods of low sunlight.

Other breeding populations

The La Mé population ( Fig. 5 ) was developed by IRHO in Ivory Coast, from 21 teneras coming mainly from Bret Plantation. Seeds from L2T, a well known line of this population, which transmit high yielding characters and short trunks, have been introduced to Malaysia, Indonesia, Costa Rica and other countries. Though the best La Mé lines were introduced to Costa Rica in 1980, progeny testing of these male parents was initiated in 1991.

A relatively unknown source of plants was prospected in the Kigoma district of Tanzania in 1978. These palms were evaluated as open-pollinated progenies of wild origin and showed excellent yield and remarkable bunch quality characteristics for unselected materials. Besides, the Kigoma lines gave the highest iodine index values among the E. guineensis origins tested by ASD.

Progeny testing of Kigoma materials was initiated in 1989. When crossed with Deli duras, the progenies showed high early yield and reduced vegetative growth. Because such materials were collected at high altitude, they are likely to be tolerant to cool temperatures.

Commercial yields of ASD Planting Material

Besides genetic potential, climatic and soil factors largely determine the oil palm production (Ferwerda 1977; Ong 1982; Ochs and Daniel 1976; Olivin 1968; Ng 1968). Also, crop management practices play an important role in current yield levels. Therefore, the production potential of any planting material in commercial plantations is always reduced to a variable extent when compared to experimental results.

Though there is quite a large area planted with ASD material in Latin America (191,000 hectares), not many plantations provide reliable and accurate long-term production records. Some plots and small plantations growing under very good environmental and management conditions have produced around 40 tons of FFB in peak years, a fact that reflects the high yielding potential of ASD Deli x AVROS material. However, due to the factors mentioned before, the yield figures obtained from large areas in several countries are normally lower and very variable, as shown in Table 5. The yield patterns are shown in Fig. 6 , Fig. 7 and Fig. 8 .

The main agroecological conditions in the regions where the above plantations are growing are the following:

Monte Plata, Dominican Republic

Located at 18o North, there is a 3-month cold and dry season. During that time, rainfall is less than 40 mm per month and minimum temperatures fall below 16°C. Soils are red-clay Latosols, shallow to moderately deep, of very low fertility, on a rolling landscape.

Quepos, Costa Rica

Located at 9° 30' North. It has a 3- to 5-month dry season, with an average annual water deficit of around 325 mm. Soils are mainly alluvial shallow to deep, silt loamy, somewhat poorly to moderately well-drained, with very high calcium and magnesium content, on flat lowland.

Coto Costa Rica

Located at 8° 30' North, this area has a mild dry season with an average annual water deficit of less than 100 mm. However, owing to heavy cloudiness, sunlight is low to very low during most of the year. Soils there are mostly alluvial, deep loamy to clayey, poorly to moderately well drained, with very high calcium content, on flat lowland. However, there is an important area with volcanic soils (Andosols), which are shallow, loamy moderately well-drained, of low fertility and high phosphorus fixing capacity, on flat to gently-sloping landscape.

Quininde, Ecuador

In this region, at 0° latitude, there is a 4- to 5-month dry and cold season. Though it rains below 50 mm per month during that time, due to persistent heavy cloudiness the average sunlight is less than 50 hours per month, and the minimum temperature drops below 17°C. The soils are volcanic (Andosols), deep, loamy, well-drained, of medium to low fertility and high phosphorus fixing capacity, on undulating land.

Tecun Uman, Guatemala

This zone, located at 14° 30' North, presents a 4- to 6-month dry season, with an average annual water deficit of around 500 mm, conditions which make oil palm cultivation possible only with irrigation. Soils are volcanic and alluvial, moderately deep to deep, well-drained, sandy loamy to light clayey, moderately fertile, on rolling and gently-sloping landscape.

Tiquisate, Guatemala

This area, located at 14° 15' North, has a 4- to 6-month dry season, with an average annual water deficit of 500 mm, which makes oil palm cultivation possible only with irrigation. Soils there are alluvial, deep, loamy to sandy loamy, well-drained, very fertile, on flat landscape.

Santa Marta, Colombia

This region, located at 11° North, has a 5- to 6-month dry season, which makes oil palm cultivation possible only with irrigation. Soils are mostly alluvial, deep, moderately well- to well-drained, loamy to clay clay-loamy, slightly alkaline, on flat landscape.

Though in all the areas described above there are important climatic and soil limitations for oil palm cultivation, the average FFB yield records of plantations with ASD materials are very good. It is also clear that limitations such as severe water deficit and cold temperatures not only suppress yield, but also cause high variability of FFB yields between years. Other limitations, such as low sunshine during some months and poor drainage, significantly depress yields. However, the impact of most of these adverse factors can be partially reduced with good management, as is the case in Tecun Uman, Guatemala, where yields up to 38.5 tons of FFB/ha/year have been achieved from irrigated 8-year old plantations.

Quality of ASD's Planting Material

When talking about oil palm seeds, quality can be assessed in two ways; the genetic and physical aspects. The breeding program must guarantee a planting material that carries all the yield characteristics expected by the planter. The physical aspects are ensured by taking the necessary precautions to offer a product which meets all phytosanitary standards and has an excellent appearance.

The significance of genetic quality involves the guarantee of purity, also referred as the legitimacy of the seeds. For ASD, this matter is of top priority, because it is closely linked with reproducing, at commercial level, the results observed in the breeding experiments. ASD seeds are guaranteed at 99.9% genetic purity. Quality is systematically evaluated while monitoring seed production. Asystematic sampling of 100 commercial hectares in 1987, produced seven duras (99.9% purity). In 1991, the evaluation of 300 hectares showed no duras, a result which was again obtained in 1995 when 400 hectares were evaluated.

As a result of maintaing high quality standards, the costs of seed production are invariably high. Pollination techniques require costly materials and manpower if the goal is to guarantee quality without cutting corners. ASD follows a standard methodology, established in 1976 in cooperation with the IRHO. Expensive pollination bags are imported to guarantee 99.9% teneras, because no reliable local substitute has been found yet.

References

Breure, C. J.; Rosenquist, E. A.; Konimor, J. and Powell, M. S. 1985. Oil palm introductions to Papua New Guinea and the formation of selection methods at Dami Oil Palm Research Station. Proceedings of International Workshop on Oil Palm Germplasm and Utilization. Selangor, Malaysia. p 189-197.

Chan, K. W.; Ong, E. C.; Tan, K. S.; Lee, C. H. and Law, I. H. 1986. The performance of oil palm genetic laboratory (OPGL) germplasm material. Proceedings of International Workshop on Oil Palm Germplasm and Utilization. Selangor, Malaysia. p 162-175.

Ferwerda, J. D., 1977. Oil palm. In: Alvim, P. and Kozlowski, T. T. Editors, Ecophysiology of Tropical Crops. Academic Press Inc., New York, USA, pp. 351-380.

Ng, S. K., 1968. Soil suitability for oil palms in West Malaysia. In: Turner, P.D. Editor, Oil Palm Developments in Malaysia. Incorporated Society of Planters, Kuala Lumpur, Malaysia. pp. 11-17.

Ochs, R. and Daniel, C., 1976. Research on techniques adapted to dry regions. In: Corley R.H.V. et al Editors, Oil Palm Research. Elsevier Scientific Publishing Company, Amsterdam, The Netherlands. pp. 315-330.

Okwuagwu, C. O. 1985. The genetic base of the NIFOR oil palm breeding program. Proceedings of International Workshop on Oil Palm Germplasm and Utilization. p 228-241.

Olivin, J., 1968. Etude pour la localisation d'un bloc industriel de palmiers a huile. Oléagineux, 23(8-9): 499-504.

Ong, H.T., 1982. System approach to the climatology of oil palm. II. Identification of temperature and sunshine aspects. Oléagineux, 37(10): 443-450.

Rajanaidu, N.; Tan, Y. P. Ong, E. C. and Lee, C. H. 1985. The performance of inter-origin commercial DxP planting material. Proceedings of International Workshop on Oil Palm Germplasm and Utilization. p 155-162.

Richardson, D.L.; 1995. The history of oil palm breeding in the United Fruit Company. ASD Oil Palm Papers, Costa Rica, N° 11: 1-22.

Rosenquist, E. A. 1985. The genetic base of oil palm breeding populations. Proceedings of International Workshop on Oil Palm Germplasm and Utilization. p 27-59.

Van der Vossen. 1974. Towards more efficient selection for oil yield in the oil palm ( Elaeis guineensis Jaq.). Center for Agricultural Publishing and Documentation, Wageningen. The Netherlands. 107 p.

Wonki-Appiah, J.B. 1974. Prediction of yield and fruit qualities in the oil palm Elaeis guineensis Jacq Ghana Journal of Agriculture Science 7:3. 209-213.