Plant Breeding: Improvement Objectives

A. Yield Potential

The oil palm is the highest yielding oil crop (Corley 1985; Robbelen 1990), its highest annual recorded oil yield of 10 t/ha is 2-5 times those of annual oilseed crops such as soybean and rapeseed. Yield improvement in the oil palm plantations for the past 50 years has been attributed to 70% cultivar improvement and 30% improvement in agronomic practices (Davidson 1993). Breeding progress in the Deli dura planting materials was about 15% per generation. (A breeding generation in oil palm requires 8-10 years; three years to bearing, four to six years of yield measurement, and one year of hybridization). With the switch-over to the tenera planting material there has been at least 30% improvement in oil yield mainly through better oil content from the thicker mesocarp (Hardon et al. 1987). Since then there have been two generations of improved tenera hybrid planting materials and oil yields have improved from 4.9 t/ha to 8.9 t/ha but this could not be attributed entirely to breeding as the materials were planted at different times and locations and perhaps with different agronomic inputs (Lee and Toh 1992).

Current planting materials derived from mixed hybrids thus possess high yield potentials. The theoretical yield potential based on crop physiological computations is 17 t/ha oil from 45 t of fresh fruit bunches (Corley 1985). Oil yields exceeding 12 t/ha, fresh fruit bunch yields of 45 t/ha, and oil content in bunch of 35% have been reported in experimental plantings, although not necessarily jointly. Nevertheless they indicate that the prospect of significant quantum increase in yield potential is rather limited, at least in the near term. For larger quantum yield improvement drastic changes with the crop architectural design (plant form and planting pattern), cultivation and harvesting methods would be needed.

Increase in crop yield potential in other crops has been achieved through the combination of high biomass production and high harvest index, the former through high density planting (Evans and Fischer 1999). With current genotypes and cultivation system (uneven terrain plantings and high dependence on manual operations particularly for harvesting) plant densities of higher than 138 to 160 palms/ha currently practiced would seem inadvisable. However, proposals and efforts are underway to breed for high efficiency palms with more efficient light capture, higher photosynthetic rate, higher leaf expansion ratio, higher leaf area ratio, better conversion efficiency of energy captured to dry matter, reduced respiratory loss and improved harvest index (Breure and Corley 1983; Squire 1984; Breure 1985, 1986; Henson 1992; Smith, 1993). Many of these contentions are based on inferences drawn from measurements made on individual palms in mixed stands and on correlations based on small sample sizes which were prone to auto-correlations or statistical dependencies because most of the physiological parameters were derived variables in which total dry matter or yield constituted a major component (Rajanaidu and Zakri 1988; Chang and Rao 1989). Proof of the value of these physiological or ideotype breeding traits (Rasmusson 1991) would have to be obtained from replications of such attributed palms within plots and between plots in experimental plantings. These are as yet not forthcoming. In cereals, higher yield potential has been achieved through higher biomass and/or better harvest index using dwarfing genes (Peng et al. 1999; Reynolds et al. 1999) while increased yield in maize has been attributed to tolerance to high density planting (Tollenaar and Wu 1999). The Dumpy semi-dwarfing gene(s) in oil palm did not contribute to better harvest index (Soh et al. 1981). Despite this, slight increases in harvest index and planting densities are conceivable with new genotypes in the breeding pipeline (Rajanaidu et al. 2000).

While the concept of genetic yield potential is helpful in achieving the physiological limit, the concept of harvestable yield is important for a perennial tree crop where the actual harvesting (cutting) operation (Plates 1B, 1C) is manual with little prospect of a feasible and cost effective mechanized alternative. Harvestable or recoverable yield relates to ease of harvesting, harvesting at optimal fruit ripeness and minimal loss of ripe fruits. Dwarf palms with longer bunch stalks will facilitate harvesting, manual or mechanical (Soh et al. 1994b). Palms with ripe fruits exhibiting a distinct color change will ensure harvesting at optimum ripeness while those with non-abscising fruit habit will minimize loose ripe fruit loss. Such programs have been proposed using breeding or genetic engineering approaches (Osborne et al. 1992; Rao 1998) as these traits are found in less advanced breeding or germplasm materials.

Plate I. B) Four year old oil palm harvested with a chisel.

Plate I. C) A 20 year old oil palm harvested with sickle attached to aluminum pole.

B. Adaptability

Much of the reported planting material improvement is likely to have been through adaptability to site limitations. This involves moving the site yield potential (Tinker 1984) towards the genetic yield potential by circumventing site limiting factors e.g. soil, moisture, terrain, rather than through improvement in genetic yield potential per se. However, genotype × environment (GxE) interaction has not been generally considered to be a serious factor (Rosenquist 1982; Cochard et al. 1993). The presence or absence of GxE depends on the specific genotypes and specific environments tested (Corley et al 1993). The misperception that GxE is not important could be attributed to the use of the genetically variable commercial sources of planting materials as the genotypes (Chan et al. 1986; Rajanaidu et al. 1986); the use of related progenies in different environments (Cochard et al. 1993) or the use of related progenies in similar environments (Rosenquist 1982; Lee and Rajanaidu 1999). Results of other experiments and analyses (Obisesan and Fatunla 1983; Obesisan and Parimoo 1985; Ong et al. 1986; Corley et al 1993) have detected the presence of GxE effects using genetically diverse progenies although their contributions (3-4%) to the total experimental variance for yield were still small (Rajanaidu et al., 1993; Rafii et al., 2000). Although purported stable progenies have been identified (Ong et al. 1986; Lee et al. 1988; Rafii et al. 2000) their exploitation as cultivars depends on the ability to reproduce these tenera hybrid progenies in large quantities, as only limited hybrid seeds can be produced from a pair of dura and pisifera parents. Most of these authors also did not indicate the characteristics that conferred stability in the progenies or attempt a biological explanation of the GxE effects (Caligari 1993; Lee and Rajanaidu 1999). Corley et al. (1993), using a simple analytical approach found that progenies with few bigger bunches rather than many smaller bunches tended to yield poorly in stressful environments because the abortion of a single large bunch would result in a considerable loss in yield. Plasticity of bunch weight was also found to vary between progenies.

In terms of selection, an attractive approach used in animal breeding to handle GxE effects is to treat a trait in different environments as separate but genetically correlated traits (Falconer and Mackay 1996). They can then be incorporated into a selection index to predict selection response for the trait in the second environment as a correlated response to selection for the trait in the first environment (White and Hodge 1989; Yamada 1993; Soh 1999).

With the advent of near true hybrid cultivars from inbred parents or cloned parents and clones, GxE effects with respect to location, spacing, fertilizer, and other factors will likely assume increased importance (Lee and Donough1993; Corley et al. 1995; Soh et al. 1995) and may warrant consideration in the development of cultivars suited for different situations, in line with the trend towards precision farming or precision plantation practices (Chew 1998).

C. Oil Quality

The oils produced by the oil palm are very versatile in their uses in the food and oleochemical industries. Despite this, palm oil has yet to penetrate the cooking and salad oil markets of the temperate countries. Its large saturated fatty acid component (Table 1) makes the oil with a melting point of 36ºC (Yusoff 2000) solidify at colder temperatures. Because of this and the fact that it is a palm-derived oil, consumers have incorrectly lumped it with coconut oil as a saturated fat unhealthy for human consumption leading to a predisposition to heart disease. This is despite the fact that palm oil contains about 50% unsaturated fatty acids and that the saturated component is mainly palmitic acid which Khosla and Sundram (1996) contended to be neutral in its cholestrolemic behavior. Palm oil also contains anti-oxidants in its carotene and tocopherol/tocotrienol contents which appear to have anti-carcinogenic properties. It can also be used directly, i.e. without hydrogenation, for margarine without the production of the unhealthy trans fatty acids.

Table 1 . Variability of fatty acid composition and iodine value in oil palm populations. Source: Arasu 1985; Rajanaidu 1990a; Rajanaidu et al 2000.

Fatty acid

Content (%)

Elaeis guineensis

E. guineensis

E. guineensis

E. oleifera

E. oleifera ×
E. guineensis














3.8 -14.7






33.0 -55.9






5.4 -15.8





Iodine value




zPalm Oil Research Institute of Malaysia.
yInstitut de Recherches pour les Huiles et Oleaigneux

Nevertheless, end-user and consumer bias has pressured breeders to breed for a more liquid palm oil. The ideal palm oil should have the following fatty acid composition based on the American Heart Association’s (1990) recommendated fat intake: total energy requirement derived from dietary fat intake should not exceed 30%; one third from saturates [lauric (C12:0), myristic (C14:0), palmitic (C16:0), stearic (C18:0)] which are hypercholestrolemic; one third from polyunsaturate [linoleic (C18:2)] which is hypocholestrolemic; and one third from monounsaturate [oleic (C18:1)] which is neutral cholestrolemic. In this respect, palm oil is excessive in saturates and deficient in polyunsaturates.

Advanced oil palm breeding populations have low genetic variability for unsaturated fatty acid contents, generally expressed as iodine value. The situation is better in the West African semi-wild prospected materials but the highest values can be found in the American oil palm, E. oleifera , collections. Palm Oil Research Institute (PORIM), currently known as Malaysian Palm Oil Board (MPOB), recommended elevating the iodine value to about 70 by breeding to put palm oil on a competitive basis with olive oil (Task Force 1985). To do this without sacrificing the current progress in yield and agronomic attributes using selections from the collected E. guineensis materials would involve 4-5 generations of backcross breeding to the advanced breeding materials. With the E. oleifera materials, more backcross generations are envisaged because of interspecific hybrid infertility (Hardon 1986, Sharma 2000). The same would apply to breeding for higher carotene and tocopherol/tocotrienol contents. Reservations to this approach have been expressed (Hardon and Corley 2000). Because of the long lag breeding time (20-30 years), by the time the cultivar is developed, the market may have changed and the perceived premiums disappeared. Genetic and agronomic manipulations are more expedient with annual crops, which can respond readily to market changes. An oil quality genetic improvement program also requires a large separate effort, which may be at the expense of the main yield improvement program through dilution of effort and selection pressure. It may be more expedient to concentrate on yield improvement and achieve higher quality component production from more oil produced and using chemical and bioprocessing technological advances.

D. Stress Tolerance

Being a profitable and easily grown crop, oil palm plantings have expanded into sub- optimal and marginal areas (e.g. dry sandy areas, podsols, peat, highlands). In these areas biotic (disease and pest) and abiotic (water, temperature, nutrient) stresses, which impede or reduce production, are likely to be encountered.

1. Biotic. Basal stem rot caused by the basidiomycete fungus Ganoderma boninense is the only disease warranting consideration in resistance breeding in the Far East. It used to be a malady of older palms planted on former coconut or oil palm areas with high water tables but reports of its occurrence on younger plantings in inland areas have become more frequent (Ariffin, 2000). Progress in resistance breeding had been hampered by the lack of an efficient screening technique (Khairudin et al. 1993; Ariffin et al. 1995). No resistant or tolerant genotype has been found to date with the preliminary screening technique developed (Ariffin, 2000). In a very recent paper (De Franqueville et al 2001) parental, progeny and clonal differences in resistance/susceptibility were demonstrated in palms planted on Ganoderma infested soil. Ganoderma disease control is the subject of an international cooperative effort coordinated by CAB International (1998) in which resistance control is an important objective.

Because a reasonably efficient nursery screening technique is available, tolerance to Fusarium (Fusarium oxysporum f. sp. elaedis) vascular wilt forms an integral part of the breeding programs in West Africa (De Franqueville and Renard 1990). Lethal bud rot debilitates oil palm in Latin America. Any proposal to breed for tolerance to this disease would be premature as the pathogenic cause of the disease is still in contention (Ariffin 2000).

Crown disease, a physiological affliction of young 2-3 year old oil palms in the Far East, causes bending and twisting of the young fronds, which can result in loss in early yields when severe and prolonged. The Deli dura material appears to be more susceptible. The disorder is caused by a recessive gene, the expression of which is masked by an epistatic gene conferring incomplete penetrance (Blaak 1970). The disorder can be bred out by discarding families with any susceptible progeny.

2. Abiotic. Oil palms in West Africa experience reasonably long periods of drought and hence West African commercial planting materials tend to be more drought tolerant because they have been bred under such conditions. There are also drought tolerance breeding programs (Houssou et al. 1987; Okwuagwu and Ataga 1999). The semi-wild collections obtained by MPOB from the drier regions of Nigeria would include drought tolerant genotypes. Palms introduced from the Bamenda Highlands of Cameroun (Blaak 1967) are presumed to possess genes for cold tolerance although their likely precocious flowering behavior under lowland conditions have been touted.

Manifestations of Mg deficiency occur on oil palms planted in sandy areas in some parts of Papua New Guinea and Indonesia. Tolerance to Mg deficiency is an objective of the breeding program in Papua New Guinea (Breure et al. 1986). Palms planted on deep peat tend to lodge. Dwarf or smaller palms would circumvent this problem. Peat plantings are also prone to micronutrient deficiencies, particularly Cu, Zn, and B. As these deficiencies can be corrected easily with micronutrient applications, a breeding approach is unnecessary.