1. INTRODUCTION

Selection in oil palm aims to increase the production of oil and kernels per ha. A selection strategy to develop elite planting material, involving large-scale testing of parent palms, has been formulated by Breure & Bos (1992). But information on the practical aspects to implement such a program is scarce.

The main economic product is palm oil extracted from the mesocarp. The shell thickness is therefore an important characteristic as this determines the proportion of the fruit available for the oil bearing mesocarp.

Shell thickness is determined by a single gene. One homozygote, the pisifera, is shell-less; many pisifera palms fail to fruit, so the pisifera is not grown for commercial use. The other homozygote, the dura, has a thick shell. The heterozygote of the dura x pisifera cross, the tenera, has a thin shell. The tenera is the fruit form preferred for commercial use, because more of the pericarp consists of mesocarp than in the dura.

As pisifera usually produce bunches with predominantly sterile fruits, the dura is used as the female and the pisifera as the male parent of tenera planting material.

The search is thus for dura and pisifera parents which transmit high bunch yield and oil-and-kernel extraction per hectare to their tenera offspring.

Breure & Bos (1992) proposed to select dura and pisifera parents in three steps:

1. preliminary selection of dura and pisifera palms on phenotypic characters, i.e. those characters measured on the parent palms,
2. further selection among these on the basis of General Combining Ability (GCA) values, i.e. the additive genotypic effects of the parents, obtained from a progeny test, and
3. testing of families, derived from intercrossing palms selected in step (ii) in all combinations, with the main objective of exploiting both GCA and Specific Combining Ability (SCA), i.e. the contribution of the interaction effect of parents on the performance of the offspring.

Although the SCA effect is usually much less than the GCA effect, the ultimate aim is to search for specific crosses between dura and pisifera parents.

Within these elite crosses one can find outstanding tenera. Clones derived from these tenera may yield higher than the source family. Testing clones may therefore be the final stage in upgrading planting material.

The three-step selection of dura and pisifera parents is explained in section 2.1. Section 2.2. outlines the procedure of estimating the contribution of GCA and SCA effects on the performance of the offspring.

Section 2.3 shows how a proper choice of the mating design can enhance the precision in comparing GCA values of pairs of parents.

Usually, a large number of parents is involved in such testing trials; progress in crossing work depends on the availability of female inflorescences on the dura parents and males on the pisifera. An efficient method of implementing the crossing program is described in section 2.4. The lay-out of progeny experiments should be such that progenies are as much as possible arranged in blocks of uniform soil conditions. However, it is usually difficult to find sufficiently large blocks to accomodate the total set of progenies because of diversity in drainage and other physical and chemical soil characteristics. In that case progenies must be arranged in incomplete blocks, i.e. by subdividing replications into smaller homogeneous blocks containing only part of the progenies, as is illustrated in section 2.5.

The statistical methods to identify the set of elite parents in step 2 for further testing in step 3 is described in section 2.6.

The way secondary characters are taken into account in the selection strategy is described in section 2.7.

The general breeding strategy is described in section 2.8.

Section 2.9 describes the size and shape of plots for experiments to select parent palms and also those to compare sources of planting material.

During the final stage of recording, differences in the progenies's ability to compete for light may create an important diversity in mutual shading among progenies; the stronger competitors then tend to gain yield at the expense of their weaker neighbouring progenies. The way the effect of light competition on selection efficiency can be minimized by the arrangement of progenies in the field is described in section 2.10.

Plantation companies often like to test planting material from distinct origins at the onset of large-scale planting programs. Section 3.1 describes the trial design for evaluating different sources of planting material; its statistical analysis is given in section 3.2.

The actual technique of recording yield and growth has often received little attention. These include yield, oil and kernel extraction, secondary selection characters derived from records to determine the area and weight of the leaves, leaf production, trunk diameter and height increment.

Frequent and accurate measurements are needed to obtain meaningful parameters. Section 4.1 covers the recording of the components of oil and kernel yield and growth illustrated with technical drawings; in particular a novel technique to measure vertical stem growth is described.

Section 4.2 covers measurements to estimate growth parameters.

Section 4.3 outlines the period of recording of the various parameters.

Section 4.4 illustrates the components of growth and section 4.5 the calculation of growth parameters.