BIODIVERSITAS ISSN: 1412-033X Volume 21, Number 8, August 2020 E-ISSN: 2085-4722 Pages: 3534-3538 DOI: 10.13057/biodiv/d210815 Genetic diversity and selection within natural dura oil palm accessions collected in Ghana for oil palm productivity improvement DICKSON OSEI DARKWAH 1, , ESSIE BLAY 2 , HARRY AMOATEY 3 , ENOCH SAPEY 1 , CLAUDE BAKOUMÉ 4 , DANIEL AGYEI-DWARKO 1 1 Oil Palm Research Institute. P.O. Box 74, Kusi-Kade, Ghana. Tel. +23-324-8556687, 4611325, +86-130-41293087,  email: oseidarkwah@yahoo.com 2 Department of Crop Science, University of Ghana. P.O. Box LG 25, Legon, Accra, Ghana 3 Ghana Atomic Energy Commission. P.O. Box LG 80, Legon, Accra, Ghana 4 Maxi Productivity Sarl. P.O. Box 2137, Douala, Cameroon Manuscript received: 22 May 2020. Revision accepted: 10 July 2020. Abstract. Darkwah DO, Blay E, Amoatey H, Sapey E, Bakoume C, Agyei-Dwarko D. 2020. Genetic diversity and selection within natural dura oil palm accessions collected in Ghana for oil palm productivity improvement. Biodiversitas 21: 3534-3538. The genetic base of oil palm breeding stock of the Ghanaian Oil Palm Research Institute is considered narrow after two selection cycles. Therefore, 79 accessions of natural dura oil palm collected in Ghana and conserved at Oil Palm Research Institute (OPRI) were field evaluated for 18 agro-morphological traits. ANOVA revealed differences among the 79 dura oil palm accessions for 14 traits. The most heritable traits were single fruit weight (h 2 = 0.85), leaf area index (h 2 = 0.63), bunch weight (h 2 = 0.52), total leaf area (h 2 = 0.51), and fresh fruit bunch (h 2 = 0.46). Genetic coefficients of variation ranged from 3.30% for frond dry weight to 68.74% for total bunch weight denoting possibility of selection of desirable traits for improvement. Five high fresh fruit bunch yielding accessions (FFB ≥ 10.81 t/ ha) with good heritable bunch characteristics were selected for introduction in the breeding stocks. In fact, their bunch number, average bunch weight, and mesocarp to fruit were higher, in absolute terms, than the controls and trial mean obtained in the study. Although, the height increment was close to those of the controls, short progenies are expected among descendants from their selfings or crosses. Keywords: Average bunch weight, fresh fruit bunch, genetic coefficient of variation, heritability, natural oil palm INTRODUCTION The oil palm (Elaeis guineensis Jacq.) is a diploid, tropical perennial monocotyledonous species indigenous to Africa. It has 2n = 2x = 36 chromosomes. It is naturally cross-pollinated but is not self-incompatible. On the basis of the shell (Sh) thickness, three types of oil palm are distinguished. Notably (i) dura with a thick shell (2-8 mm) and homozygous for the trait (Sh + Sh + ), (ii) pisifera shell- less and homozygous (Sh - Sh - ), and (iii) tenera a hybrid from the cross between the dura and pisifera (Sh + Sh - ) characterized by a thin shell (0.5-2 mm). Following the discovery of the genetic inheritance of shell thickness character (Beirnaert and Vanderweijen 1941) since the early1960s, tenera was adopted worldwide as planting material (Bakoumé et al. 2017). Oil palm is cultivated for its oils including palm oil contained in the mesocarp and palm kernel oil contained in the kernel. Palm oil and palm kernel oil have food and non-food uses. Increase in human population and diverse uses of the crop led to a high demand for palm oil that could not be satisfied by natural groves (Bakoumé et al. 2016). Oil palm selection and breeding work commenced early in the last century in most major oil palm research centers of the world's humid tropics including OPRI in Ghana. The aims of oil palm breeding in Africa are to maximize palm oil (and kernel) yield, to reduce vertical growth to extend the plant's economic life, to select for tolerance to drought to enable expansion of the planting area, to explore new sources of tolerance to Fusarium wilt, and most recently to improve palm oil quality notably high iodine value and low free fatty acids content (Bakoumé et al. 2016). Unfortunately, breeding programs relied on restricted genetic base populations (Ataga et al. 2012) and resulted in slow genetic progress from one cycle of selection to the next, which was the case of Ghana. Varieties of commercial lines produced from such narrow genetic base will be highly similar genetically. In fact, planting materials are tenera (D x P) i.e., hybrids derived from a cross between dura (D) female parents and a pisifera (P) pollen donors. Pisifera is female sterile, the flowers abort after fertilization. Dura is generally Deli dura derived from only four palms introduced in 1848 in the Bongor Botanic Gardens in Java, Indonesia. Dura is known to transmit tolerance to diseases, vertical growth rate, bunch production, and all cytoplasmic-controlled genes to the tenera (hybrid) progenies (Bakoume and Louise 2007). The narrow genetic base of Ghanaian dura breeding stocks has resulted in difficulty of selection for desirable traits. Therefore, there was a need to re-vitalize OPRI's breeding programs by enlarging and enriching the genetic bases of its dura breeding stocks to guarantee substantial genetic advance through selection. Okyere Boateng et al (2008) indicated that oil palm landraces possessing adaptation to local environments are gradually being displaced through (i) felling of palm trees for traditional palm wine production, locally distilled into alcohol, (ii) introduction of improved types, (iii) settlement on new lands due to