Abstract

Bunch analysis records from 1990 to 1995 for nine Deli dura x AVROS progenies from a progeny test trial located in southwestern Costa Rica showed a cyclical behavior in the oil extraction rate and the kernel to bunch ratio. It was noted that the variance observed in the variable "oil to bunch" was caused mainly by changes in pollination efficiency (fruit set) and, to a lesser degree, by the variance shown by the variables "oil to mesocarp" and "mesocarp to fruit".

The differences recorded in the variable "fruit to bunch" in a single progeny during different periods of the same year showed the cyclical effect of climate on pollination efficiency, particularly due to the reduced production and viability of pollen, in addition to the reduced size of the pollinating insect population which was usually seen between September and November. These changes explained the low proportion of fruit to bunch seen in mature bunches harvested between February and May. Variation in precipitation and sunlight a few weeks before harvest coincided respectively with changes in the variable "oil to mesocarp" in bunches harvested between February and May, and between July and September.

These results showed that the ranking of progenies done according to data obtained during the fifth year after planting has a closer relation than any other year in the study, which included data from year three to year seven. Furthermore, in year five, the three most important determining variables for oil to bunch (oil to mesocarp, mesocarp to fruit, and fruit to bunch) had the smallest variance.

During the fifth year, a minimum of 80 and a maximum of 151 bunches, in oil palm progenies represented by at least 36 palms, should be analyzed in order to obtain a faithful representation of the quality parameters of the bunches.

Introduction

The oil extraction rate and the kernel extraction rate are commonly-used to measure the efficiency of an oil palm operation. While these parameters are useful indicators of operational costs, they do not show the actual productivity level of the plantation, and are therefore capable of measuring neither the efficiency of the harvest nor other limiting environmental factors that prevent the achievement of optimum yield.

On a commercial plantation, productivity depends on environmental conditions, the yield potential of the genetic material planted, and the administrative and agronomical efficiency.

Optimum oil and kernel yields depend on bunch production and quality, vegetative growth which determines the commercial life of the plantation, and the ability of the crop to overcome inter--palm competition and other detrimental external factors like disease and water deficit.

Due to the high degree of heredability of some oil palm bunch components, the improvement of bunch quality has been used as a tool to increase yield potential. Various authors (Rao et al ., 1983; Link & Toh, P. 1988) have mentioned disadvantages in this standard method, among which they included the introduction of a systematic positive bias during the taking of samples, from which subsamples are later taken.

Another disadvantage commonly associated with this method is the high degree of variability shown by some bunch components, which necessitates a substantial number of analyses to obtain a suitable progeny characterization.  This factor also significantly increases the number of samples needed to describe and select individual palms.

This paper reviews sources of variation that affect the main selection parameters related to oil palm bunches. The effects of these sources of variation on oil and kernel extraction rates are analyzed. In addition, the minimum sample size and evaluation period in order to accurately describe and select progenies using bunch characteristics as selection criteria are determined.

Materials and methods

In September, 1987, nine Deli dura (BM8 x BM20) progenies crossed with AVROS (BM119/7) pisiferas were planted in the experimental station of the ASD Palm Research Program (Programa de Investigación de Palma de ASD), located in southwestern part of Costa Rica. The area is characterized by deep alluvial soils, with good fertility, and by low sunlight and high precipitation (3800 mm/year).

On a monthly basis, 25 mature bunches from each progeny were collected and analyzed (Blaak et al . 1963). The study began in August, 1990, 39 months after field planting, and continued until June, 1996. Bunches were collected when they showed a natural detachment of three to five fruits.

Using the data obtained between September, 1990 and August, 1995, the annual variance for each variable was estimated and minimum and maximum values were obtained. The variables under study were "average fruit weight", "average bunch weight", "total fruit per bunch", "fertile fruit per bunch", "mesocarp to fruit", "oil to mesocarp", "kernel to fruit" and "oil to bunch".

Both short--term and mid--term tendencies were studied for the variables. The progenies were classified every year for the five years of the study, using as a criterion the magnitude of the variable "oil per palm per year" (OPAY).  Using the information for the entire period, a final ranking was made of the progenies using the variable OPAY, which was then used to calculate the values of the "Spearman rho" indices, which were, in turn, used to compare the different rankings of progenies made during the individual years of the study.

In order to determine the principal sources of variation, the statistical procedure known as "stepwise regression analysis", (SAS) was used, using "oil to bunch" as the dependent variable and the other bunch components as the independent variables.

The sample size for the fifth year was obtained by using individual sampling units for the variable "oil per palm per year"; a confidence of 95% with a precision of 5% around the mean was achieved. The initial population size was established as 12 bunches per palm and 36 individuals per progeny, and a average sex ratio of 0.8 was assumed. Thus, the standard sample for a progeny in the fifth year was 346 bunches.

Results

Palm age, climate and bunch components

Both yield and average bunch weight increased with palm age, this tendency was maintained until palms reached nine years of age ( Fig. 1 and Fig. 2 ). The variance of "fruit size" showed few changes with palm age ( Table 1 ). However, there was a considerable seasonal effect related to climatic changes ( Table 2 ).

The average value of the variable "fruit to bunch" decreased with palm age and also showed significant variation in response to seasonal climate changes. The proportion of mesocarp to fruit was inversely related to "fruit to bunch". This second variable showed a decrease as palm age increased, and, in addition, it was negatively related to the variable "kernel to fruit"

Oil content in the mesocarp consistently rose beginning in the third year, reaching a maximum around the fifth year. A slight seasonal effect, was related to climatic changes.

"Oil to bunch ratio" gradually increased until the fifth year ( Table 1 ). The main determinants of the variance in this parameter were, in order of importance: "oil to mesocarp", "mesocarp to fruit," fertile fruit to bunch," and the palm age.

Sources of variation and their influence in determining sample size

During year five, the lowest values for "oil to mesocarp" and " oil to bunch" were observed ( Fig.3 ). Total proportion of fruit to bunch" and "fertile fruit" showed their highest point in year four and declined from there ( Fig.4 ). The variables "kernel to fruit" and "mesocarp to fertile fruit" had a consistent variation in years three, four and five, and a significant increase in year six ( Fig. 5 ).

The variance in "average bunch weight" showed a consistent increase beginning with the harvesting of the first bunches and lasting through year six. Then, in year seven, it decreased ( Fig. 6 ), while "fruit weight" had its peak in year four and a valley in year six. There was no specific period or year in which the variances were the lowest simultaneously for all the variables under consideration.

The variance in "average bunch weight" showed a consistent increase beginning with the harvesting of the first bunches and lasting through year six. Then, in year seven, it decreased ( Fig. 6 ), while "fruit weight" had its peak in year four and a valley in year six. There was no specific period or year in which the variances were the lowest simultaneously for all the variables under consideration.

The variance in "average bunch weight" showed a consistent increase beginning with the harvesting of the first bunches and lasting through year six. Then, in year seven, it decreased ( Fig. 6 ), while "fruit weight" had its peak in year four and a valley in year six. There was no specific period or year in which the variances were the lowest simultaneously for all the variables under consideration.

The regression analysis showed that 49.1% of the variance in "oil to bunch" was due to the variance of "oil in the mesocarp", 37.9% was due to "total fruit to bunch", and 12.5% to "mesocarp to fruit". The variable "oil to mesocarp" showed a valley in year five, "total fruit per bunch" showed its peak in year four, and "mesocarp to fruit" in year six. Thus, the smallest variations for the three most important variable in the determination of "oil content in the bunch" are seen in year five.

The Spearman correlation test showed that, more than any other year, year five had the progeny classification that corresponded most closely to the ranking of the progenies done with data analyzed and grouped for all seven years (Table 3).

The sample size in year five, which permited a valid description of the principal bunch components using a sampling framework of 346 bunches under the conditions described, was: 151 bunches for the progeny with the highest variation, associated with the variable "oil production per palm per year"; 99 bunches when the study was done on a control progeny; and 80 bunches for the progeny with the lowest variation. The result of the sampling above was a confidence of 95% and precision around the mean of 5%.

Discussion

It has been documented that the oil and kernel extraction rates are subject to cyclical variations caused mainly by seasonal climatic changes and changes in palm age (Sparnaaij 1962; Henson 1993).

Oil content in the mesocarp is determined principally by the degree of bunch ripeness and by other external factors related to hereditary variation caused by the genetic origin of the material planted, as well as by climatic factors like sunlight and water deficit in the weeks prior to bunch ripening (Azis 1985; Siregar 1976; Henson 1993). In this study, the oil content in the mesocarp decreased in April and September.  The first reduction can be linked to the water stress caused by the accumulated water deficit of 280 mm which is generally seen in the first trimester of the year in this area. The second drop in oil content in the mesocarp is associated with the reduction in sunlight which regularly occurs between July and September. Finally, palm age is yet another factor which significantly affects the oil content in the bunch.

The variable "fruit to bunch" showed a large variation, both between individual palms and in the corresponding data for one single palm, with values ranging from 40% to 84%. The differences in values in periods of the same year suggest differences in pollination efficiency, which can be explained by both the scarcity of viable pollen and the reduction in the size of pollinating insect populations between September and November,(Alvarado y Sterling 1996; Chinchilla and Richardson 1989). The changes noted are reflected in the extraction rates registered between February and May, when the bunches harvested showed a low proportion of fertile fruit. This same period of poor pollination, extrapolated at the moment of harvest, coincided with the low values seen in the variables "kernel to fruit" and "mesocarp to fruit" between February and May.

In addition, the proportion of fruit to bunch went down and bunch weight went up with palm age. This behavior, previously explained by Lim & Toh (1985) and Henson (1993), is determined that in the bunches with the greatest weights generally have more spikeles, as well as more flowers per spikelet.

Finally, due to the fact that the greatest stability in the variance of the main determining parameters for the variable "oil production per palm per year" is noted in year five, this period must be included in studies of oil palm progeny selection. These results showed that the ordering of progenies done with data obtained in the fifth year after planting had the closest relation to the ranking done using data from the entire period of study (year three to year seven).

References

Alvarado. A.; Sterling. F. 1996. Variación estacional en la tasa de extracción de aceite en el cultivo de la palma aceitera. ASD Oil Palm Papers (submitted)

Aziz, A.A. 1985. The biochemichal aspects of ripeness standard. In Proc. Of the Symp. on impact of the pollinating weevil on the Malaysian oil palm industry., PORIM. Kuala Lumpur. p. 165--176.

Blaak, G. et al . 1963. Breeding and inheritance in the oil palm Elaeis guineensis Jacq. Part II. Methods of bunch quality analysis. J.W Afr. Inst. Oil Palm Res., 4:146--155.

Chinchilla. C.M.; Richardson. D.L. 1989. Situación actual de los insectos polinizadores y la polinización de la palma aceitera. Boletin Técnico OPO--United Brands 3 (2): 29--48.

Harbor, A. and Runyon, R. Estadística General. Fondo Educativo Interamericano, Méjico, 1973. p. 214--217,324.

Henson, Y.E. 1993. Factors determining mesocarp oil to bunch ratio in the oil palm: A physiological perspective. Proc. Nat. Sem. Palm Oil Extr. Rate: Problems & Issues. PORIM, Malaysia. Pp. 27--35.

Lim,  K.H.; Toh, P.Y. 1985. The accuracy and precision of bunch analysis. Proc. of. Symp. on impact of pollinating weevil on the Malaysian oil pam industry. p. 91--109.

Rao. V. et al . 1983. A critical reexamination of the method of bunch quality analysis in oil palm breeding. Inst. Pengelidikan Minyak Kelapa Sawit Malaysia (PORIM, Kuala Lumpur. Occasional paper No.9, 28 p.

Siregar, I.M. 1976. Assessment of ripeness and crop quality control . In Proc.. Malay. Int. Agric. Oil Palm Conf. Kuala Lumpur. p. 711-715.

Sparnaaij, L.D., 1969. Oil palm. ( Elaeis guineensis ) In: F.P. Fewerda & F. Wit (eds.) Outlines of perennial crop breeding in the tropics. Misc. papers Agric. Wageningen. 4:339--387.

Tam, Tai Kim. 1975. The relationship between the degree of fruit set and bunch components of fruit bunches of the oil palm ( Elaeis guineensis JACQ). Planter, Kuala Lumpur  51: 541-547.