Studies on deterioration and aflatoxin contamination in copra during storage

ARTICLE

Introduction

Copra is the product obtained after drying coconut kernel, from an initial moisture content of approximately 50 % to a moisture content which is ideally close to 6 %. Unfortunately, the drying process is frequently unsatisfactory, and if the copra is under-dried this can lead initially to bacterial damage, and later to attack by moulds and insects.

The qualitative effects of deterioration in copra have been known for many years. Work prior to 1937 has been reviewed by Cooke [1], while Child [2], reporting in 1974, provides a useful summary of the problem. The quantitative effects of deterioration have been examined by Nathanael [3] who emphasised that simple measurement of oil content and free fatty acid (FFA) before and after storage does not give an indication of true storage loss. If copra is considered to be composed of water, oil and proteinaceous meal, then deterioration may involve a differential loss of both oil and meal. If, during fungal attack, the loss of meal is greater than the oil loss, then the stored copra may show an oil content higher than that of good quality copra originating from the same source. Indeed, some oil millers prefer mouldy copra because of the higher oil recovery [4]. Nathanael estimated the loss copra during the storage of under-dried copra (at 12% moisture) over a 5 month period to be 15%, of which 10% was avoidable. The FFA of the oil after this period was 1.5% compared with 0.8% in well dried copra stored over a similar period.

In recent years an additional problem associated with copra deterioration has been recognised, namely the presence of aflatoxin in copra that has been attacked by the mould Aspergillus flavus. When copra is processed in an expeller, aflatoxin in the copra is distributed between the oil and cake [5]. Since the main use for copra cake is in the manufacture of cattle feeds, the presence of aflatoxin in feeds for lactating cattle gives particular cause for concern because of the known metabolism of aflatoxin B₁ in copra cake to aflatoxin M₁ in milk [6]. This is reflected in European Community regulations [7] which limit the level of aflatoxin B₁ in raw material imports to 20 ppb (µg/kg) and limits levels in dairy feeds to 5ppb. Meeting this limit requires the promotion of improvements in copra quality, for example through a grading scheme. However, any scheme must provide a real financial incentive to cover the increased processing costs necessary for the production of dry, mould-free copra [8].

Copra marketing in the Philippines

Most of the copra produced in the Philippines is made by small-holder farmers. The majority is initially traded by weight under the traditional "Pasa" system, in which the prevailing area price of copra is adjusted according to its estimated moisture content and other attributes. The marketing system for copra in the Philippines has two major functions.

(a) to buy and accumulate copra in large quantities required by the oil mills, many of which have capacities in excess of 200 tonnes per day;

(b) to allow time for the material to dry to a level at which it can be processed by the factory for oil extraction.

The storage period between production and processing generally varies between four and eight weeks. It is during this period, when the copra slowly dries from 15 to 16% to around 6%, that significant damage from both mould and insect attack is likely to occur. However, during periods of market speculation, storage periods can be considerably longer. The trading system is complex and may involve several different routes [9], but in general the copra resides in a minimum of three stores before processing, those of the primary trader, the dealer and the factory.

Purchases by the dealer, and at the factory, involve moisture content determination using the Brown-Duvel method. This technique gives results some 2 to 3 percentage points higher than the oven method [10]. Much of the copra from the farmer has a moisture content well above the level of 6% (oven method) recommended for safe storage. Estimates of the average initial moisture content by Brown-Duvel are as high as 18% (i.e. 15 to 16% oven). When trading copra, it is usual to express moisture content on a "wet weight basis", and all figures recorded in this paper are presented in this form.

The work described in this paper was carried at the Davao Research Center of the Philippine Coconut Authority (PCA) in Mindanao. The trials were designed to provide information on :

(a) conditions which promote deterioration in copra and the development of aflatoxin ;

(b) the effect of the method of copra drying on its stability ;

(c) the relationship between the appearance of copra and its quality ;

(d) the likely level of post-harvest losses in the Philippine coconut industry.

Trial outline

Materials

Quality changes were monitored in copra stored at different moisture contents and produced by three drying methods

Drying was carried out on half cups. The timing of each trial was carefully planned so as to avoid overloading the analytical facilities.

The variety of coconut used for these trials was the "Laguna Tall". Sun dried copra was dried directly on the ground for a maximum period of 5 days. The kukum dryer is fired by coconut husk and uses oil drums as the heat exchanger. The tapahan is the traditional direct heat "smoke" dryer, which also uses coconut husk as fuel. Changes in the quality of commercial copra from various parts of the marketing chain were examined. These were:

(d) Sun dried deliveries to a primary trader (ref. PS1 to PS5)

(e) Tapahan dried deliveries to a primary trader (ref. PT1 to PT5)

(f) Copra delivered to a large dealer (ref. BD1 to BD5, drying method unknown)

(g) Copra being delivered to a large factory (ref. BF1 to BF5, drying method unknown)

Sample preparation

The copra was stored in bags to allow recovery of the sample from the store. The bags (used locally for onions or potatoes) had an open mesh which allowed for easy transfer of moisture, mould spores and insects during the storage period. Individual bags contained between 6 and 10 kg. With the exception of the commercial samples, the copra for each storage set was selected to be of the same moisture content. The chosen moisture contents were approximately 24%, 18%, 12% and 6%. These were judged from the known period of drying and the texture of the copra. The preparation of the 16 bags for the first trial with the kukum dryer (K1) at 24% moisture content was as follows:

Thus the storage trial with non-commercial copra required 16 bags of copra for any given initial moisture content, (i.e. 64 bags for the four moisture contents selected). Commercial copra was well mixed before bagging the replicate samples. As in the above example 16 bags were prepared for each sample taken. The complete study involved the preparation of 896 bags of copra of which 672 were placed into store, the remainder being the controls (storage time 0 week).

Storage conditions

The initial concept was to bury the prepared bags of copra in large stacks of copra to allow equilibration to the environmental conditions within the commercial stack. However in practice, such bags were extremely difficult to recover, and deep storage was replaced by shallow storage in which the bags were kept at the surface of the stack covered with 15-30cm of copra.

Storage conditions were selected to approximate to commercial practice. Bags of copra with high moisture levels (24% & 18% ) were initially buried within a stack of copra in a primary store. When the trader sold his stock, the copra was moved to a dealer's store and buried again. Copra designated "12%" moisture was buried in a dealer's store, and low moisture copra buried directly in a factory store. All copra kept for more than four weeks was finally stored at a factory in a heap of copra that was not required for immediate processing.

Analytical methodology

Control samples of copra (dried overnight at 60°C) were analysed on the day when the copra was placed in store. Stored materials were analysed at 2, 4, and 8 weekly intervals. Measurements were carried out to determine the following

* Visual appearance by separating the copra into three types based on the extent of mould attack

* Moisture loss over the storage period.

* Changes in oil quality (FFA and colour).

* Aflatoxin B₁ levels.

* An assessment of dry matter (oil plus meal, oil loss and dust formation.

Visual appearance of the copra

The copra was separated into three types based on the extent of mould attack

(a) Satisfactory copra, substantially free from mould attack.

(b) Copra with slight or superficial mould attack.

(c) Copra with severe mould attack i.e. attack resulting in surface penetration. It was sometimes necessary to break the copra to see this penetration.

Copra with any sign of mould penetration was designated "severe" while the differentiation between copra placed in the "satisfactory" and "slight mould" groups was less rigid. All three fractions were weighed and expressed as a percentage of the total weight.

In addition the presence of moulds and insects was recorded together, where possible, with their identification. Dust present in the sample was separated and weighed.

Moisture loss

The relatively large samples of copra employed in these trials (10 to 20kg for each duplicate) were handled by the following method designed to overcome sampling difficulties and possible excessive weight loss during grinding. This two stage process involved drying the bulk sample at 60°C, followed by drying a ground sub-sample at 103 ± 2°C.

The bags of copra were weighed and pre-dried in a large fan-assisted, electrically heated, tray dryer overnight at 60°C. After cooling, the copra was reweighed and the two bags comprising each replicate were combined and ground in a knife mill fitted with an 8mm screen. The copra was then passed through a sample divider and a one twelfth sample removed. This sample was further mixed by hand prior to removing sub-samples for the oven determination in which the moisture content was determined as the weight loss when the ground copra (approximately 5g) was heated for three hours at 103 ± 2°C. The weight loss of the bulk sample together with oven method determination was used to compute the moisture content of the original sample.

Oil content of copra

The analysis of oil content was carried out by NMR in a Mark 111 Newport quantity analyser using crude coconut oil as standard. The method requires a dry sample, and the combined duplicates from the oven moisture determination were used. The results for several samples were compared with those given by the Soxhlet method; agreement was satisfactory.

Isolation and analysis of the coconut oil from each treatment set

The residual ground dry copra after sampling for moisture and oil content (9.2 to 18.4kg) was processed in a small-scale expeller (throughput 40 to 60 kg/hour with an extraction efficiency of 70 to 80%) to provide crude oil and copra cake. A small sample (100ml) of the freshly prepared oil was clarified by filtration through a plain folded Whatman number 1 filter paper. Oil colour was measured by the Philippine Coconut Authority standard method on a Lovibond Tintometer using a 5.25 inch cell. The FFA of the oil was determined by titration with 0.1N sodium hydroxide solution using a method based on a British standard (BS 684: section 2.10: 1988). All recorded levels of FFA are presented in terms of lauric acid.

Aflatoxin analysis

Aflatoxin analysis was conducted out on samples of cake from the expelling process rather than the copra itself. The cake was ground in a knife mill to pass a 2mm screen and sub-samples (approximately 2 x 50g) were obtained using a sample divider. These were analysed for aflatoxin using an HPTLC method combined with automatic fluorescence scanning for detection and quantification [11].

Oil and dry matter loss

Losses of dry matter and oil were calculated by relating samples on a "weight into store" basis. The success of this assessment relies heavily on the initial selection of copra with homogeneous moisture content within each storage set.

Aflatoxin in expeller cake

A separate study [5] established that the ratio of aflatoxin in the cake produced by the small-scale expeller to that in the copra from which it was derived was approximately 1.5: 1. In addition, the ratio of aflatoxin in the cake, from copra processed in commercial expellers in the Philippines did not exceed 1: 1. Aflatoxin data presented in tables 1 and 2, figures 1, 2 and 3, and plates 1 to 6 inclusive are calculated from the aflatoxin in copra cake x 0.66. These can be interpreted as being the level of aflatoxin in the copra, or the maximum that is likely to be found in cake after commercial processing in an expeller.

Results

Visual appearance

The control samples for each series of sun-dried samples suffered mould attack during preparation and the extent of this attack increased with the drying period. However, all the attack was superficial. Black moulds on the sun-dried copra were identified as Curvularia. The kukum dryer generally produced white copra but the under-dried samples proved to be highly vulnerable to mould attack. Tapahan copra was mould free but tended to be dark in colour due to smoke deposition. All the high-moisture samples suffered "severe" attack by moulds within the first two weeks of storage, the most prevalent being Rhizopus sp. (white), A. niger (black),A. flavus (yellow-green) and A.Glaucus (green). Some were found to contain copra beetles (Necrobia rufipes). At four weeks several samples were infested with flour beetles (Tribolium castaneum) and mould mycelia was found to have broken down the copra structure, forming dust and leaving a stained surface. In the case of the high moisture samples, the copra surface was badly pitted. Sometimes copra that had earlier suffered superficial attack was classified as "satisfactory". Copra with very "severe" damage was often fragile and disintegrated easily.

The relationship between the proportion of cups by weight showing attack by penetrating moulds, and their initial moisture content is shown in figure 4. This is an important relationship, since the extent of mould attack should also be related to oil quality and aflatoxin content. The scatter of data about the "fitted" curve in figure 4 suggests that copra dried in a tapahan to a moisture content lower than 15% may be less prone to mould attack than copra prepared in a kukum dryer, or sun-dried to a similar moisture level.

The visual quality of the sun-dried copra from trial S3 (initial moisture content nominally 6%, 12% and 24%) after eight weeks storage is shown in plates 1 to 3 respectively.

Moisture loss

The moisture contents of all the copra samples examined fell to 5-7% within the initial 2 week storage period (figure 5a, b and c). A separate study (figure 6) showed that copra with moisture contents as high as 20% will dry (given good ventilation) to safe moisture levels within 10 days. However, as indicated above, it is during this time that moulds are likely to develop. In relation to the Philippine marketing system this approximates to the period of on-farm storage, and drying in the primary traders' store.

Quality of extracted oil - Free fatty acid development

Changes in FFA content in relation to initial moisture content are presented in figure 7 for the eight weeks storage period. The data indicate that the extent of FFA formation is related not only to the moisture content into store, but also to the method of copra preparation. This is perhaps not surprising because

(a) sun drying was carried out under very poor conditions (intermittent sunshine with periods of rain) and the copra had deteriorated before it entered the store. The average FFA of oil from this control was 0.7% as compared with 0.1 to 0.2% for the copra without deterioration.

(b) The average FFA in oil from the control kukum and tapahan copra was 0.2% and 0.1% respectively. Tapahan copra differs primarily from copra prepared by sun drying or in a kukum dryer due to smoke deposition on the inner surface. In addition there is generally some case hardening on the copra surface in the smoke-dried copra due to the high temperatures involved.

Estimates of the average increase in FFA during 2, 4 & 8 weeks storage intervals are given in table 3. Sun dried copra demonstrated the highest increases in FFA. FFA increases found in the kukum and tapahan copra are of the same order up to an initial moisture content of 16%. Copra with good initial visual quality after preparation in either a kukum or tapahan dryer, when placed in a store at 10% moisture would be expected to yield an oil with a level of FFA of less than 1% after eight weeks storage. Although not evaluated in this trial, sun-dried copra of equivalent quality would be expected to exhibit similar storage properties. The relationship between FFA increase and the percentage of cups by weight with penetrating mould is shown in figure 8. As a general observation, oil present in mouldy copra will continue to deteriorate after the copra has dried out, irrespective of the method of preparation, although the rate of deterioration may decline as the product ages (figure 8).

Oil quality - Colour

Oil colours at the end of eight weeks storage are summarised in table 4. Oil from well prepared kukum or sun-dried copra is water white. However, in these trials poorly sun-dried copra produced oils with a high red colour. Such copra was characterised by the presence of black surface mould (Curvulavia sp.). Oils from sun-dried copra were pink to deep red in colour. They would not normally be observed within industry because the copra for processing is usually of mixed origin. There are indications that the red colour in oils extracted from poor quality, sun-dried copra diminishes with long storage. In general mould attack leads to an increased yellow colour in oil. The colour of oils from copra prepared by sundrying, in the tapahan and in the kukum dryer and stored for eight weeks is shown in plates 4 to 6 respectively. With the exception of the oils from samples S3G18 & S3G24 (plate 4) all meet the Philippine standard for crude coconut oil (maximum 12 red and 75 yellow).

Aflatoxin development

Levels of aflatoxin B₁ for the three different methods of manufacture are shown graphically in figure 1. The relatively high levels of aflatoxin found in some of the controls was the result of problems in handling in the earlier part of the trials. There was little change in the level of aflatoxin after two weeks storage and on average, aflatoxin levels slightly declined on storage. On the basis of these observations the aflatoxin levels in the stored copra for any one batch is taken as the mean for the 2,4 & 8 weeks periods.

The relationships between the initial moisture content of copra into the store and aflatoxin B₁ development for each type of copra is shown in figure 2. There is a strong similarity between the behaviour of kukum and sun-dried copra although for any single moisture content the former is likely to develop a higher level of aflatoxin. It was observed that copra from a kukum drier appears more likely to exhibit a wide range of moisture contents due to uneven drying [13] with the result that material with a low average moisture can be attacked by A. flavus. Moisture in sun-dried copra is likely to be more evenly distributed if well managed and turned regularly. Figure 2 clearly shows that copra prepared in a tapahan is less likely to contain high levels of aflatoxin.

An analysis of the data, using a Statgraphics software package, indicated that the relationship between aflatoxin levels and "into-store" moisture content (MC) are as follows

Aflatoxin B₁ (ppb) = C x % MC ^2.326

where C = 0.170, 0.123 and 0.047 for copra prepared in a kukum dryer, by sun drying and in a tapahan dryer respectively.

Therefore to achieve an average level of 20 ppb aflatoxin B₁ (as required by current EC regulations), kukum and sun-dried copra should be dried to 7.8% and 8.9% respectively, while tapahan dried copra only has to be dried to 13.5%. However, if the average level of aflatoxin is 20ppb then around 50% of the samples will exceed this figure. To reduce the potential incidence of aflatoxin contamination to 5% of samples it will be necessary to dry to average marker levels of 4.6%, 5.3% and 8.8% for kukum, sun-dried and tapahan copra respectively. The very low figures for the kukum and sun-dried copra reflect the necessity for long drying periods in order to overcome the effect of uneven drying.

The relationship between aflatoxin and the percentage by weight of cups showing penetrating mould attack after eight weeks storage is presented in figure 3 for the three drying methods. It can be seen that for any given level of attack, tapahan prepared copra is likely to contain lower levels of aflatoxin than that produced in a kukum dryer or by sun drying. It is clear that any sample of sun or kukum dried copra showing more than 20% attack by penetrating moulds is likely to exceed an aflatoxin B₁ level of 20 ppb. The relationship between aflatoxin levels and mould attack was calculated to be

Aflatoxin B₁ (ppb) = e ^{(0.035 x % mould + C)}

where C = 2.36, 2.96 & 1.57 for copra prepared by sun, kukum and tapahan drying respectively.

Predicted average values for aflatoxin B₁ (ppb) in relation to penetrating mould attack are as follows :

Material losses

Losses of dry matter and oil over the eight week storage period for the non-commercial samples are presented graphically in figures 9 and 10 respectively. Linear regression analyses gives the following relationships :

% dry matter loss = 0.28 x % moisture - 1.51 R = 0.76

% oil loss = 0.25 x % moisture - 2.09 R = 0.70

Estimates of average material losses for copra kept under the trial storage conditions are recorded in table 5. The figures confirm the known stability of copra held under good conditions of storage with a low initial moisture content (5 to 7%). However, there is no evidence of an anticipated rise in the oil content of the deteriorated copra although this is known to occur in practice. The mean oil content of the "control" copra in these trials was 66.7% (dry basis) while the calculated oil content after storage (based on the data in table 5) is 66.9% over the moisture range 10 to 20%. Measurement of dust formation was obscured by the fact that dust was lost during the movement of the bags but could also be gained within the store from surrounding copra. No relationship between the initial moisture content and dust formation was observed. Average levels of dust formation for the non-commercial samples over an eight week storage period were 0.3%, 0.4% and 0.3% for the sun, kukum and tapahan dried copra respectively.

Commercial samples

Quality changes in the commercial samples after eight weeks storage are summarised in table 1.

The commercial samples taken from the dealer and factory stores showed a wide range in quality. Oil FFA after 8 weeks storage ranged from 1.6% to 4.6% and aflatoxin levels from 12 to 592 ppb. As the age of the samples was unknown, only the level of severe mould attack can be applied as an indicator of aflatoxin contamination (table 2). In nine of the twenty samples examined the aflatoxin B₁ content found was much higher than that predicted.

Discussion

The study confirms the long established fact that copra containing 5-7% moisture, which is largely free from mould attack, is a stable commodity. When under-dried copra is placed into store, attack by A.flavus will result in build up of aflatoxin B₁ until the copra dries out to a moisture content of approximately 8%, below which the mould is no longer active [2]. Given good storage conditions this occurs over the first two weeks of storage, which in the Philippines is in the early part of the marketing chain, and approximates to the period on farm and in the primary trader store. The study highlights the protective effect against aflatoxin contamination imparted to copra by smoke drying in a tapahan. This compares well with observations by de Lava in 1928 [13] and more recently by Arseculeratne and his co-workers [14]. Arseculeratne studied the effect of smoke from coconut shell and found an inhibitory effect on aflatoxin accumulation rather than mould growth. Such inhibition has been observed in the present study (figure 3) but there are indications that tapahan drying actually inhibits mould formation (figure 4). Whether the latter is the result of a component of the smoke, case hardening or a combination of these two factors requires further study.

However, use of the direct fired tapahan introduces an additional quality problem, contamination by polycyclic aromatic hydrocarbons (PAH), although it is known that PAH can be removed from coconut oil if the refining process includes treatment with suitable levels of activated carbon [15].

The study indicates that copra entering the marketing system with an average moisture content of 16% and stored with good ventilation will suffer a material loss of less than 3%. In practice copra tends to be stored in large heaps and material at the centre could well suffer increased loss.

The results also provide baseline data for the development of grading schemes to limit average aflatoxin levels to 20ppb. Ideally copra should be produced at source at 6% moisture as quickly as possible (generally a minimum of 5 days sundrying or 24hrs in a kiln). However, a moisture acceptance limitation at the first point of sale to 9% should provide a satisfactory result. At this level, after drying in the store to around 5-7%, tapahan prepared copra is highly likely to meet the aflatoxin limit while sun-dried copra should average 20ppb. However, copra prepared in a kukum dryer would require drying to an average moisture content of 8%. This can be attributed to the wide moisture distribution associated with this natural draught dryer, a problem overcome in the tapahan by the protective action of the smoke. Further down the marketing chain when the copra has dried out, visual quality becomes important, and here the percentage of cups showing attack by penetrating moulds is a guiding factor.

Relating visual mould attack to aflatoxin levels can be difficult. A batch of very mouldy copra may contain little A.flavus and hence a low level of aflatoxin. Another batch of what may appear to be good copra, may contain small pockets of A.flavus in amounts sufficient to create significant contamination.

In a separate study some 120 pieces of copra, each contaminated with A.flavus were combined and found to contain 1180 ppb of aflatoxin B₁. The average surface cover of mould was 40%. At this level of contamination one cup in fifty eight (1.7%) is sufficient to result in a mixed sample exceeding 20 ppb of aflatoxin. Restricting mould levels in tapahan and sun-dried copra to a level below 20% should, in theory, eventually yield copra cake with an average level of aflatoxin B₁ of 20 ppb. However, in practice, observed aflatoxin levels in some samples are much higher than predicted so lower mould levels, say 5%, will be required to ensure compliance. Any measure taken to control aflatoxin levels will also improve oil quality. Using the measures described in this paper, FFA levels in coconut oil prepared from copra stored for 8 weeks would be expected to be no greater than 1%.

CONCLUSION

Acknowledgements

The authors would like to acknowledge the practical assistance of Manuel de Castro, Rose Balucan, Evelyn Tan, Lorna Maximo, Florante Glacio, Teodoro Bontillao and Carlito Gestopa of the Philippine Coconut Authority and Sarah Phillips, Clifton Gay and Phillip Drew of the Natural Resources Institute. Aflatoxin analyses were carried out by the PCA in Manila under the direction of Nelly Valencia. The authors are indebted to John Wood for his valuable assistance with the preparation of the manuscript.

The study could not have been carried out without the ready cooperation of the following organisations : The Davao Research Center (PCA), Mintal Commercial , Solid Asia Marketing Corporation, Julu Enterprises and Legaspi Oil Mill who are all located at Davao City Mindandao. The work constituted part of an RP-UK technical cooperation project designed to reduce aflatoxin levels in copra cake/meal.

The Department for International Development of the United Kingdom can accept no responsibility for any information or views expressed in this presentation.

REFERENCES

1. COOKE FC (1937). The practical aspects of copra deterioration. Dept Agric 55 and FMS, Bull 28 (General series).

2. CHILD R (1974). Coconuts 2^nd ed. Longman : 247-9.

3. NATHANAEL WRN (1960). Economic losses to the coconut industry consequent on deterioration of under-dried copra. Ceylon Coconut Q, 11 (3/4) : 5-46.

4. FERNANDEZ W. (1987). Oil loss in copra due to fungi: a review. The Philippine Agriculturist 70, 1 and 2 : 29-40.

5. HEAD SW, NAGLER MJ, HARRIS RV (1993). An assessment of aflatoxin transfer from copra to products during the manufacture of crude coconut oil. Presented at a CIRAD-CP seminar "European research: working for coconut" in Montpellier, Sept 8-10th.

6. ALLCROFT R, CORNAGHON RBA (1962). Vet Rec, 74 : 863.

7. European Commission (1998). Commission Regulation (EC) n° 1525/98 of 16 July 1998, amending Regulation (EC) n° 194/97 of 31 January 1997 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Communities L201:436.

8. GRIMWOOD BE (1975). Coconut palm products. FAO Agricultural Development Paper, 99 : 110.

9. MANUEL P, MAUNAHAN, (1983). Coconut/copra marketing system-Structure, conduct and performance. Los Banos: Philippine Council for Agricultural Resources Research and Development.

10. HEAD SW, HARRIS RV (1995). Examination of the Brown-Duvel method for the determination of moisture in copra. Coconuts Today, October.

11. TOMLINS KI, JEWERS K, COHER RD, NAGLER MJ (1989). A bidirectional HPTLC development method for the detection of low levels of aflatoxin in maize extracts. Chromatographia, 27 (1/2) : 49-52.

12. BAWALAN DD, HEAD SW (1993). Evaluation of the traditional and an improved kukum dryer for copra manufacture. Unpublished work.

13. DE LAVA VG (1928). Chemical studies on soconut products 1. The critical moulding - moisture content of copra and some methods of preserving it. Philippine Agriculturist, 15.

14. ARSECULERATNE SN, SAMARAJEEWA U, WELIANGA LV (1976). Inhibition of aflatoxin contamination in smoked substances. Journal of Applied Bacteriologyu 41 : 223-33.

15. LARSSON BK, ERIKSSON AT, CERVENKA M (1987). Polycyclic aromatic hydrocarbons in crude and deodorized vegetable oils. JAOCS 64(3) : 365-70.