Initial research on oil palm in Nigeria started in the period between 1930 and 1950, when this country was one of the most important supplier of palm oil in the world. Several research stations were created where important steps were taken to develop outstanding genotypes now used in several breeding stations in Ghana and in Costa Rica. This paper traces back the origin of many exceptional palms evaluated in Nigeria, and whose descendants served as the source of some commercial materials now sold by ASD of Costa Rica. Oil palm seeds where received in Costa Rica from Nigeria in 1977, and since then two commercial varieties have been developed, and others will be released in the future.

First attempts to develop an oil palm research program in Nigeria come from 1930, given the importance this country had played in the supply of palm oil for world trade up to WWI. However, due to WWII, The West African Institute for Oil Palm Research (WAIFOR) did not start its functions until 1952. The main station (1735 ha) was located 32 km from Benin, and a substation was established at Abak in the Calabar province (182 ha). Some Agriculture Department stations utilized for the program were the Moor Plantation (Ibadan), Ogba (Benin), Nkwele (Onitsha), Umudike (Umahia) and Obio-Akpa (Abak). The institute became eventually The Nigerian Institute for Oil Palm Research (NIFOR).

From 1911 to 1915 E. Smith of the Nigerian Department of Agriculture planted a plot of about 800 oil palms at Calabar in southeastern Nigeria. Yield and bunch composition of these plants (coming from open pollinated seeds collected from local wild grove palms) was collected from 1922 to 1928. From this work, nine duras and ten teneras were selected. Twelve of the selected palms were self-pollinated to form the Calabar F1 generation and the progenies were planted in four stations of the Department of Agriculture between 1930 and 1935 (Ogba, Umudike, Ibadan and Nkwelle). These stations supplied extension work seed (EWS) and seeds for experimentation from these Calabar F1 s.

More selfs of the original Calabar-selected palms were made for planting at the Benin main station in 1942. The palm 551.256D (Ca 256) is the best known of the original Calabar palms for its outstanding yield. The occurrence of virescens fruit color in Nigeria Calabar materials traces to 551.341 and 551.375 which were among the selected palms.

The Ogba station produced DxD, TxT and DxP crosses for planting at the Benin main station in 1945-47. This was the Calabar F2 generation seed. Thus began the strange practice of keeping the population of duras separated from the teneras in the WAIFOR program, while the different origins were mixed.

From the Nkwelle station Calabar F1's, 14 palms were selected, and from the Umudike station 16 Calabar palms were selected for further breeding. Selection at both stations was based on fruit yield and bunch composition. The selected mother palms were pollinated by bulked pollen from the other selected palms. Again the dura and tenera groups were kept separated. These progenies were planted in a Calabar "concentration area" at the Benin main station in 1958.

Hence by 1958 the main station had planted the following Calabar germplasm:

1942: F1 progenies from the original Calabar plot

1945-47: F2 progenies from the Ogba station

1940's: Calabar DxP extension work seeds from the outlying stations 1

958: F2 progenies from bulked pollen of Calabar selections at Umudike and Nkwelle stations planted in the "concentration area".

In the 1920's a palm grove improvement experiment was initiated in a 11 ha grove near Aba. Yield and bunch composition of about 200 palms were recorded during 12 years. By the late 1930's five Aba dura palms and six teneras were selected. These palms were selfed and intercrossed and the progenies were planted at the Benin main station in 1939-42. Since the pollinaton techniques at Aba were poor, the resulting progenies were largely the result of outcrosses. Considering this shortcoming, attention centered on TxT progenies from four selected Aba teneras, while DxD progenies were of little interest.

Later additional bunch composition data were collected on the Aba palms and additional selections were made. Probably only 42 duras and three teneras were selected, and a crossing program for these palms was initiated in 1952. Pollen from selected palms was bulked and used, keeping the dura group separate from the tenera group. Forty-nine Aba progenies were planted in a replicated trial at the main station in 1956, and other Aba progenies were planted in unreplicated trials.

Yield recording of wild groves palms (with a high percentage of teneras) in a 48 ha grove near Ufuma began in 1939. Based on fruit yield and bunch composition, 25 dura and 35 tenera palms were selected for use in breeding. Unlike the case in the Calabar and Aba populations, more selection pressure could be exerted on bunch composition in the Ufuma population because of higher quality bunches. Hartley (1977) gives yield and bunch characters for 25 Ufuma teneras which average F/B (56.4%), S/F (14.5%) and K/F (11.7%).

Progenies were then prepared using bulked pollen from the best palms of their respective groups. F1 progenies consisted of 72 crosses using 19 selected duras were planted at the main station in a replicated trial in 1956. Similarly, 81 progenies from 25 teneras seed parents were also planted in partially replicated plots at the main station. Okwuagwu (1976) and Sparnaaij (1958) when referring to these Aba and Ufuma progenies planted at the main station refer to a 40 ha "concentration area" of the two types. Presumably this is the same area as the progeny trials mentioned by West (1976).

Although the Angola germplasm is not indigenous, the WAIFOR/NIFOR program introduced, evaluated and widely distributed a small population of Angola germplasm. This material originated from six Angola palms planted at Njala, Sierra Leone in 1926. A selfed progeny of one of these palms 907.263T (Ang 263) was planted at the Benin main station in 1941 and additional material from the six Njala palms were planted in 1960-63.

Some of these selected Angolan palms frequently occurring in pedigrees of advanced generations were: 1.2209D, 1.2215P, 1.2224T, 1.2227D, 1.2229T.

In 1926 two Deli seed lots were received in Nigeria, one from the SOCFIN Tandjong-Gentung plantation and one from a self of the AVROS Deli palm 8. This later palm had an average yield of 270 kg FFB/year. The Deli seed lots were planted at the Moor Plantation at Ibadan, Ogba and Umudike in 1926-29. Yield recording started in 1933, and the five highest yielding palms were selected and self-pollinated. The selfed progenies were planted at Ibadan in 1939.

In 1939 Deli duras from the Serdang Central Experimental Station were received in Nigeria. Ten crosses of high yielding Delis from the Serdang Avenue palms were planted in a replicated trial at the Benin main station in 1941 along with the best second generation Ibadan Deli and four African progenies as testers. In general the progeny of Serdang 19 x 65 gave good yield, but the rest of the Delis were disappointing in productivity.

In 1961 additional Deli germplasm was obtained from IRHO, Pamol-Nigeria (Cowan), United Plantations and Jamaica. Since Lancetilla (Honduras) supplied Jamaica with Deli seeds in the 30's, it is possible that the Jamaica Deli germplasm came from Sumatra and Java through Lancetilla.

The Nigerian breeding program was initiated with germplasm of poor bunch composition, since Aba and Calabar origins formed the major proportion of the material utilized. Table 1 compares three germplasm origins for genotypic values for bunch characteristics in DxT crosses. The African origins are especially poor in fruit to bunch ratios. Note that the variability among palms of the same origin can be large, however.

In addition to starting with material with poor bunch composition the performance of the Nigerian program with respect to bunch analysis was deficient. In spite of publishing the definitive paper on bunch analysis technique in 1963, the bunch analysis laboratory worked mainly on projects other than the breeding program, and wasted considerable effort testing bunch composition during the palm's first year of bearing. In consequence there were almost no meaningful oil analyses in the NIFOR breeding program up to 1972. It is to be expected, therefore, that much germplasm from the NIFOR program is at an unselected level for oil to mesocarp. For fruit quality selection the emphasis was placed on shell to fruit, since both mesocarp and kernels were considered equally useful materials. Separate groups of progenies were kept as high kernel duras or high mesocarp duras.

Yield was emphasized in selection, but growth characters apparently were largely ignored, except for selection of short stemmed palms for the "Short stemmed program". The high leaf area of progeny of palm 32.3005T was noticed, but apparently leaf area was not measured routinely. Since Hardon and Corley's work on growth and physiological parameters was not published until the early 1970's, the lack of selection criteria for these characters is not surprising.

Calabar. In the initial stages of the WAIFOR breeding program the Calabar germplasm was the most important, especially on the tenera side. As the breeding program evolved the Calabar materials became comparatively less important and emphasis was placed on the Aba germplasm. Some of the most important selected Calabar palms are indicated in Table 2 .

One of the best known of the Calabar palms was 551.256 sometimes called Ca 256. This palm was selected from among the 449 original palms at Calabar in the 1920's. This outstanding palm had a high sex ratio and produced a high number of bunches. The self of 551.256 was used as a standard cross early in the program. Yields of the self apparently suffered little inbreeding depression for the yield level was 10 per cent better than EWS. The self of 551.256 outyielded unselected bush seed by 25 per cent in an average of seven trials planted at the main station.

This palm had drought tolerance since it maintained high yield under dry conditions as indicated as follows:

This famous palm contributed to ASD germplasm in Coto (Costa Rica) in the ancestry of GHA605, GHA610 and GHA611, GHA647 thru palm 1.3056T which is on West's (1976) list of NIFOR palms with the best combining ability as shown below.

Figure 1 summarizes the original relationships between the genetic materials developed in Nigeria and those brought to Costa Rica. In Table 3 information on some parental palms in the main program is given. Palm 1.3056T was low in F/B, but otherwise had good M/F.

Another outstanding Calabar palm was 32.364T which was selected from EWS. Calabar DxP seed proved to be an important source of elite teneras in the Nigerian program. Palm 32.364T produced large bunches with very good mesocarp to fruit ratio. The selfed progeny of 32.364T was outstanding in yield and bunch composition and is Fusarium wilt tolerant. Palm 32.364T has been called "really the only good one" when referring to Calabar palms. This palm is included in the list of good combiners. The Aba x Calabar cross 6.594D x 32.364T gave excellent yields.

Palm 32.2612T was also selected from EWS. The palm had a high bunch number, good yield and good mesocarp to fruit, and was included among the good combiners. The cross between 32.2612T and 32.364T produced palm NIF 46.387T which was an elite palm for the second phase of the MBP. The Coto program in Costa Rica received pollen of 46.387T which was used on Ekona and Nigerian females.

Palm 551.340T was a second, less distinguished member of the original Calabar palms. A so called "good" palm, 103.3T was the result of a self pollination of Ca 340 and this combined with a Ca 256 descendant to give rise to palm 1.3056T which is an important component of Coto Calabar germplasm. Ca 340 also gave rise to 1.3379T by virtue of two successive self pollinations. Palm 1.3379T is involved in the ancestory of GHA607 and GHA608, hence 551.340T is well represented in Coto.

Palm 551.703D also one of the original Calabar plot palms, gave rise to palm 3.3164T by two successive selfs. Palm 3.3164T appears frequently in the Nigerian program. Table 1 shows that it has good fruit to bunch but relatively poor mesocarp to fruit. This is partly due to the high K/F percentage. In our program palm 3.3164T contributes to three accessions.

Palm 32.3005T was also selected from EWS. Selfed progeny of this palm was Fusarium wilt tolerant. Although 32.3005T was widely used in Nigerian crossing programs it was not singled out as exceptional in Nigerian reports. It was, however, included by West in the list of the best combiners.

The use of palm 32.3005T in crosses with Chemara BPRO Deli duras proved exceptional in the Sabah breeding program. Similar results with selfed descendants of this palm in crosses with Deli dura have been obtained in Coto. Palm 32.3005T is found in the ancestry of GHA605, 607, 608, 610, 611, 647 and 648 and as such dominates the Coto Calabar origin. High leaf area for 32.3005T was reported from Nigeria. In Coto, descendants of this palm have a high frond index, by virtue of modest leaf dry weight (PxS) and very broad leaflets.

Among the Coto oil palm germplasm, the Calabar origin is by far the most important. In West's list of best combiners all Calabar palms except 3.361D and 3.417D are represented in Coto. On the other hand among the Aba materials on West's list, no palm has any descendants in the Coto program.

Aba. The more important Aba palms are listed in Table 4 , but these palms are not well represented in the Nigerian materials transferred to Coto. Aba palm 39.235D is found in the pedigree of two accessions GHA608 and 609. This palm was selected from a cross between original Aba palms 631.621 x 631.37. The palm 39.235D had a very high yield of three times the plot mean yield, but very poor bunch composition. In mesocarp to bunch palm 39.235D is the worst palm listed in Table 1 .

Palm 14.758T which is involved in Coto progenies GHA606, 607 and 609 is from an Aba DxP cross, 1.81D x 1.2742A P. The Coto germplasm with the highest percentage of Aba origin is GHA609 (50% Aba, 25% Calabar, 25% Ufuma) and GHA606 (50% Aba, 50% Calabar).

The Coto Aba material is scarce and not very highly touted. Some good Aba materials not found in Coto are described below.

Palm 1.53D was selected from a cross between original Aba palms 631.931T x 631.864T. This palm was widely used in Aba crosses and appears on the list of best combiners. Palm 1.53D is a parent of two additional palms (4.1811T and 4.3488T) which were also among the best combiners. Pisiferas derived from 4.1811T were highly susceptible to Fusarium wilt.

Palm 14.525T was outstanding in yield and bunch composition ( Table 1 ). It had high fruit to bunch and was included as a "best combiner". The palm was also selected for the second phase of the MBP.

Ufuma. The Coto collection has progenies descended from al least one Ufuma palm, namely 15.4382T. This palm has been singled out for comment in several Nigerian reports. In 1963 it was added to the group of parental palms for the first stage of the main breeding program. 15.4382 had good fruit composition with exceptionally high second and third year yields. This tenera was the third best producer in the 1956, 100 acre concentration area planting.

Okwuagwu (1986) has extolled the virtues of the Ufuma palms, mainly those coming from TxT crosses, identified by its high productivity and good bunch characteristics.

Of the Coto germplasm 15.4382T is a grand parent of GHA608, one of the most promising lines of Nigerian origin in our collection.

Angola. The Coto germplasm collection has a considerable number of Angola progenies, but these have all descended from a self of a single Angola palm. Palms 1.2209D and 1.2229T were included in West's list of best combiners. Siblings of these two palms, 1.2224T and 1.2227D, were not included on the list. Angola germplasm in the Coto program traces to palms not indicated as best combiners. Next is shown the bunch composition of the four Angola palms.

The Angola material was described in Nigerian reports as having a stiff, upright growth habit. This has also been observed in Coto, especially in C9166 which is derived from a Deli x Angola cross (CHE137:253D x G851.148D). These palms have long and ridged rachises borne in a near vertical orientation.

The Angola germplasm may be of interest in increasing the genetic diversity of the seed parents in oil palm breeding. Nevertheless, the Angola material is characterized by rapid stem growth and bunch composition inferior to highly selected Delis, hence it may be difficult to incorporate into the selection program.

Deli. By 1954 the Delis in Nigeria were characterized by high bunch and fruit weights, high fruit/bunch ratio (low stalk weight), tendency for low oil to mesocarp and more flowers per spikelet.

According to WAIFOR reports Deli bunches tended to have a more open structure. It was suggested that this could allow for a better development of mesocarp tissue on the inner fruits.

An important number of Deli palms were also considered in NIFOR program; they came from several research centres including IRHO, and others were from Serdang Avenue, or through Sumatra. Some Deli palms were outstanding in M/B.

In Deli trial 5-1 planted in 1941, 10 Malayan Deli progenies and one Ibadan Deli progeny were compared with four African testers as follows:

Group 1946-58 yield (kg/palm/year) Mean of all Deli's 45.4 Serdang 19 x 65 Deli progenies 65.2 Ibadan Deli progeny 46.0 Mean of 4 African progenies 56.8 Serdang Delis 19 x 65 were by far the best of the Delis. Palms 5.368D, 5.642D, 5.1295D, 5.1225D and 5.2153D from this source were signaled as the best Delis and were included on the list of best combiners, except for 5.642. 5.368D was one of the highest yielding Deli duras. 5.1225D has been called the best of the 19 x 65 Serdang Delis with high yield and good bunch composition.

The breeding policy was to keep the Deli duras pure for seed production. Delis were not included in the MBP, but were treated as a separate program. Nevertheless, Nigerian duras were included in the MBP; hence no decision to use Deli seed parents exclusively had been made. The Deli program for phase two included the following palms or progenies:

5.1225D           (Serdang 19 x 65)

G144D             (UP, Malaysia)

5.12 x P522D   (Serdang x Pobe)

5.368 x L404D (Serdang x La Me)

Others. Table 5 and Table 6 list Nigerian germplasm arising from inter-origin crosses. Most of the palms shown had been selected for phase two of the main breeding program. None of these palms are represented in the Coto Nigerian germplasm.

The Coto collection contains other Nigerian germplasm whose origin is obscure. In December, 1976 we obtained seven seed lots from Highlands Estate of which four were Nigerian x Nigerian crosses. Only one of these seed lots germinated. This was designated HIL539 from cross EWS81/11 x NIF22.

In 1949-50 Warren Breck made oil palm seed collections in West Africa, which were planted in Lancetilla in 1951. Some of these palms were crossed with sibs and the seeds were planted in Coto in 1968 and 1969. These Nigerian materials were poorly evaluated and little used. The Nigerian materials are as follows:

The selected Nigerian germplasm was placed in the main breeding program. Initially this consisted of Calabar, Aba and Ufuma materials with one Angola palm included. By 1959 the first group of 16 selected palms were planted in the Main Breeding Program (MBP). From time to time additional palms were included and occasionally a palm was dropped from the program. Starting about 1975 new palms were selected for the second stage of the MBP. These were mostly progeny of the selected palms from the first stage. Figure 2 shows the Nigerian palms selected at various stages in the Nigerian and Ghana programs and indicates the relationship of this material to that received in the Coto program.

Experiments in Nigeria . Table 3 shows the phenotypic yield and bunch characters of the initial group of selected palms for the MBP. The yield data are virtually meaningless and the bunch quality of the selected palms was poor. A large percentage of the palms have low fruit to bunch percentage and according to the shell content, the fruit is mediocre even though this was a primary selection criterion. The palms are incompletely tested for oil to mesocarp ratios, but those available suggest considerable variation in this character.

The defects in oil testing in the early stages of the Nigerian program allowed the selection of a palm such as 1.3340D, with only 37.5 per cent oil to mesocarp.

In 1969 NIFOR published accumulated yields from progeny trials. In order to determine the yield potential of parental palms, the high yield and low yield crosses in each experiment were listed. The parental palms that repeatedly appeared in one group or the other are listed below.

Table 7 shows the comparative performance of pairs of Nigerian palms in crosses with a uniform parent in Nigerian progeny tests. The yield results are not consistent among the pairs of crosses; hence a ranking for yield can not be done with any certainty.

In terms of relative number of bunches produced and the relative single bunch weight, the pattern is much clearer and the progression of bunch number/bunch weight types is as follows:

HBN/LBW: 32.2612, 1.3379, 14.758, 1.3056, 32.364

LBN/HBW: 32.3005

The classification is valid in a relative fashion within the group of six palms under consideration.

Table 8 shows the genotypic values for bunch characters of Coto Nigerian origin palms. Palm 32.3005T is the poorest of the group for low mesocarp and high shell. 32.364T and 32.2612T have high shell and kernel. Palm 15.4382T is good in mesocarp and shell percentages, but the fruit to bunch ratio is low. Palm 39.235D has very bad bunch composition indeed.

Experiments in Ghana. Since Ghana was part of British West Africa, Nigeria had the responsibility of assisting this country in the development of oil palm. Accordingly, crosses were made in Nigeria for planting progenies of breeding materials at Kade, Ghana. The first progenies were planted in 1961 and by 1964 the germplasm transfer was completed.

Table 9 gives phenotypic characters of Ghana palms which were used as parents to produce the progenies destined for Coto. In general these had excellent bunch composition especially in terms of low shell and high oil to mesocarp.

Some progeny test data from Ghana were obtained which included tests of some immediate ancestors of Coto palms. The yield results compared progenies from crosses made on a uniform female parent of Deli or Angolan origin.

Kg FFB/plot of 16 palms

In the only direct comparison on an Angolan female palm 86T outperformed 253T progeny by 11 per cent more yield. Palm 86T as a male parent was only consistently outperformed in Ghana trials by 326T, a Yangambi type from IRHO to Nigeria. Progenies of palms 53T and 54T usually produced lower yields than 86T.

Unfortunately no growth or bunch quality data were obtained from these Ghana progeny trials. However, judging from fruit yields alone, palms G851.86T and G851.326T were the most interesting in the progeny test.

Eight seed batches from Ghana were received in Costa Rica in 1977:

Pollen from 46.387T was received in September 1977, and at the end of that year pollen from the following palms arrived:

Later, in April and August 1979 some more pollen from these palms was received along with Calabar 851.86T (1.3056T x 32.3005T). In May, some more seeds were received:

All these introductions from Nigeria were evaluated in several experiments planted between 1979 and 1982. For some, data were taken up to 1992 after planting new progenies derived from the first ones. Some results are given.

Experiment CB7909. Eight introductions were evaluated in this experiment. Based on the behavior of the teneras, the families GHA608 and GHA610 were the best in term of oil yield. GHA608 was a high FFB yielder with a very attractive bunch composition, particularly F/B and O/B. Its stem growth rate was reduced and foliar length was below the group mean (Table 10).

Looking at Ghana origins, progeny of G851:53T was superior to G852:835T, and G853:1161T was similar to G851:53T. Besides, selfings from G851:86T were negatively affected by endogamy, but this did not occur in G851:54T.

GHA608, originally from a Nigeria compound, had 50% Calabar genes, 25% Ufuma and 25% Aba. Since these were special palms in terms of vegetative growth, the best tenera were chosen and planted in 1990.

Experiment CB8101. Three very different progenies were planted in this experiment: two families have a Calabar origin and the other (GHA646), an Angola origin. Among the first, GHA648, showed an excellent bunch yield and short leaves, but it was a taller plant than GHA647. Besides, GHA648 was superior in terms of bunch quality with larger ratios of M/F, O/M and O/B. In general, fruit set was low in this experiment which negatively affected total oil production.

Angola dura palms were good in terms of bunch production, but were also very vigorous (high stem growth) with leaves of medium size. Bunch quality was poor, particularly low F/B and medium O/M, which produced a low O/B ( Table 11 ).

Experiment CB8203. Some families planted in this experiment (C9013, C9004, C9015 and C9113) were superior in terms of oil production. Among the Ghana palms, 851.253T was superior to 851.326T and 46.387T was better than 851.253T. However, F/B was low, but in some progenies like 9113, O/B and M/F were good. Among the dura descendants, the palm C9018 was outstanding in terms of bunch production and quality ( Table 12 ).

Experiments CB90.03 and CB90.04. Six tenera palms GHA608 descendants were combined with a particular Deli dura palm. All six tenera palms had good bunch characteristics and yield, particularly 45T and 178T, but low O/B, which were similar among all progenies. However, 45T was superior to its siblings in terms of OHAY due to a better FFB, and 187T had a high M/F.

Some F1 plants originated from these teneras were planted in experiment CB90.04. Oil extraction rate was superior in the teneras derived from the selfings of 187T and 178T, but bunch yield was similar among progenies ( Table 13 ).

Experiment CB92.01. Progenies derived from GHA648 had a high FFB yield when compared with Deli x AVROS. Besides this, these progenies had a reduced stem height and leaf length, and a very good bunch composition ( Table 14 ).

Several Angola progenies (GHA646) and Deli/Angola (C9018) were evaluated in three experiments combined with several males sources. In general, the descendants had a uniform bunch yield and very high oil extraction rate. From these results, a new generation was produced from 1998 to be used in special seed-production programs and to improve the Deli population ( Table 15 ).

An important finding from the experiments in Coto was the high uniformity of some of these genotypes. The lines GHA648 and GHA608 were particularly good yielders, and were chosen to be used to develop new populations.

The oil palm breeding program of WAIFOR/NIFORT allowed the commercial exploitation of a very particularoil palm population. The outstanding characteristics of those palms have been verified in several breeding stations where the germplasm has been extensively evaluated. The ASD' Oil Palm Breeding Program in Coto, Costa Rica obtained a broad sample of these germplasm which, after 20 years, has originated two commercial varieties, but still new genes wait to be used in current breeding programs.

BLAAK, G.; SPARNAAIJ, L. D.; MENENDEZ, T. (1963). Breeding and inheritance in the oil palm (E laeis guineensis Jacq.). Part II. Methods of bunch quality analysis. J. W. Afr. Inst. Oil Palm Res., 4:146-155.

 BROEKMANS, A.F.M.; TOOVEY, F.W. 1955. The Deli palm in West Africa. J. W. Afr. Inst. Oil Palm Res. 1, (3), 51.

GUNN, J.S. 1957. WAIFOR Fifth Annual Report. 1956-57. Pp 82-94.

HARTLEY, C.W.S. 1977. The Oil Palm, 2nd Ed., Longman, p. 265.

CORLEY, R.H.V.; HARDON, J.J., TAN, G.Y. 1971. Analysis of growth of the oil palm ( Elaeis guineensis Jacq.). I. Estimation of growth parameters and application in breeding. Euphytica, 20, 307-315.

MENENDEZ, T. 1965. NIFOR First Annual Report. 1964-65. Pp 59-73.

MENENDEZ, T.; BLAAK, G. 1964. WAIFOR Twelfth Annual Report. 1963-64. Pp. 49-75.

NIFOR Fourth Annual Report. 1967-68. Pp 38-77.

OKWUAGWU, C.O. 1986. The genetic base of the NIFOR oil palm breeding program. In Proc. Int. Workshop Oil Palm Germplasm and Utilization. Kuala Lumpur. PORIM. 10:228-237.

SMITH, E.H.G. 1929. The oil palm ( Elaeis guineensis ) at Calabar. 8th Ann. Bull. Dep. Agr., Nigeria.

SPARNAAIJ, L.D. 1958. Oil palm breeding and selection in Belgian Congo. WAIFOR mimeographed report.

VAN DER VOSSEN, H.A.M. 1974. Towards more efficient selection for oil yield in the oil palm ( Elaeis guineensis Jacq.). Agr. Res. Rpts., 823. Wageningen, Holland. ( or Thesis, University of Wageningen. 107 p)

WAIFOR Ninth Annual Report, 1960-61. Pp 69-77.

WONKYI-APPIAH, J.B. 1974. Prediction of yield and fruit qualities in the oil palm Elaeis guineensis Jacq. Ghana J. Agr. Sci. 7(3):209-213.

WEST, M.J. 1976. The analysis of yield data of the NIFOR oil palm main breeding programme and the choice of new parental material. Supplementary report to the Min. of Overseas Development, Mimeograph.

WEST, M.J. et al. 1976. The inheritance of yield and fruit and bunch composition characters in the oil palm - an analysis of the NIFOR main breeding program. Int. Agric. Oil Palm Conference, Kuala Lumpur, 1976.