Coconut Research Institute Repository

Fernando, S.C.

Weerakoon, L.K.

Karunaratne, S.M.

Santha, E.S.

cocos (2002-2002), 1445-52
Printed in Sri Lanka
A COST EFFECTIVE MEDIUM FOR PRODUCING EMBRYO CULTURED
COCONUT PLANTS
S C Fernando, L K Weerakoon, S M Karunaratne* and E S Santha
Coconut Research Institute, Lunuwila, Sri Lanka
* University of Queensland, Brisbane, Australia
ABSTRACT
Zygotic embryo culture of coconut is a useful research tool employed
in various fields of crop science. In the present study, the feasibility of
reducing the production cost of embryo-cu ltured coconut plants was
explored . In separate experiments, Analar grade potassium chloride,
sucrose and double distilled water in the Y3 medium were substituted with
Muriate of Potash (fertilizer grade potassium chloride), table sugar and tap
water, respectively. Furthermore, embryos were cultured in a growth
medium devoid of hormones (BAP and 2,4-0).
The results indicated that low cost ingredients can be used in culture
media without affecting the germination and subsequent growth of coconut
embryos. The results also revealed that incorporation of hormones to the
culture medium is not essential. The findings will contribute to reducing the
cost of producing coconut plants in vitro.
INTRODUCTION
Zygotic embryo culture of coconut (Cocos nucifera L.) is a useful tool
for research in various fields, such as germplasm collection, exchange and
storage (Assy-Bah et al., 1987; Karunaratne, 1988), embryo rescue (De
Guzman, 1970; Balaga and De Guzman, 1971; Karunaratne, 1987), disease
transmission (Orense and Pacumbabba, 1988), screening for disease
resistance (Rillo et al., 1988) and drought tolerance (Karunaratne et al.,
1991 ).The first published research on coconut zygotic embryo culture was
carried out by Cutter and Wilson (1954). Since then much work has been
carried out by a number of scientists to improve in vitro culture of coconut
zygotic embryos and the technique has been perfected. Accord ing to our
knowledge, in most of the investigations on zygotic embryo culture of
coconut, culture media contained analytical grade chemicals, which are
costly . Low cost ingredients such as fertilizer grade chemicals and table
sugar have been used in several tissue culture systems to reduce the
production cost (Ganapathi et al., 1995; Purnima, 1997; Sujatha and
Chandran, 1997). In coconut embryo culture, the possible use of tap water
instead of glass distilled water (Areza et al., 1995) and refined table sugar
instead of Analar grade sucrose (E Rillo personal communication) in the
culture medium, was reported recently. These findings encouraged further
45
work on cost reduction in producing embryo cultured coconut plants. The
aim of the present study was to determine the feasibility of reducing the
production cost of embryo cultured coconut plants by using low cost
ingredients and eliminating hormones in the growth medium.
MATERIALS AND METHODS
Mature zygotic embryos were collected from freshly harvested nuts
(12-13 months postanthesis) of Sri Lanka Tall coconut. The embryos were
ste rilized , cultured and incubated as described by Karunaratne et al. (1985).
In separate experiments, Analar grade potassium chloride (KCI), sucrose
and double distilled water in Y3 medium (Eeuwens, 1978) were substituted
with Muriate of Potash (fertilizer grade KCI), table sugar and tap water
respectively. All the other chemicals used for media preparation were of
Analar grade. The possibility of eliminating hormones (BAP and 2,4-0) in
the growth medium was also tested by growing embryos in the growth
medium devoid of hormones. The standard growth medium containing
Analar grade chemicals and hormones was used as the control.
During the course of germination and development of the cultured
embryos, the numbers of germinated embryos, developing into plants,
photosynthetic leaves per plant, primary roots per plant and plant height
were recorded.
Eight replicates each consisting of 20-25 embryos were used for
each treatment. A completely randomized design was used in all the
experiments. The data was analyzed using the analysis of variance. The
percentage data was analyzed using inverse sine transformation.
RESULTS AND DISCUSSION
The results indicated that quality of KCI, sucrose and water (Tables 1
& 2) did not affect the germination and subsequent growth of coconut
embryos. It is note worthy that although commercial grade chemicals were
used in similar quantities as analytical grade, it did not have any negative
effect on embryo germination and plant growth indicating that any impurities
present in commercial grade material had no deleterious effect. Reduction
of production cost of many crops including Melia azedarach L. (Sujatha and
Chandran, 1997), Brassica campestris (Okuno et aI., 1996), Musa spp.
(Ganapathi et aI., 1995) and Zingiber officinale (Sharma and Singh, 1995)
has been achieved by substituting Analar grade sucrose with table sugar.
Furthermore, Malemngahha et al. (1994) have tested a medium prepared by
using a commercial fertilizer formulation, table sugar and coconut water for
Dendrobium spp. embryo culture. The successful use of tap water instead of
46
distilled water in micropropagation of Musa spp. (Ganapathi et al., 1995),
Solanum tuberosum (Purnima, 1997) and Cocos nucifera .(Areza et al.,
1995) has also been reported . Thus the above studies have shown that
commercial grade chemicals could be used successfully in in vitro culture
and the production cost of in vitro-raised plants could be reduced
significantly. The results of the present study are in agreement with these
find ings.
The use of non-ionic water instead of double distilled water and
refined table sugar instead of Analar grade sucrose has resulted in slight
decrease in secondary metabolites content of cell cultures (Zheng et el.,
1982; Jianping et a/., 1996). However, the decrease was not significant
when compared to the cost reduction achieved by using cheaper chemicals.
The results of the present study revealed that coconut embryos
could be germinated and grown successfully in media devoid of hormones.
Furthermore, the number of leaves per plant developed in hormone-free
medium was significantly higher than that of plants developed in the control
medium (Table 3). This would be beneficial as plants having more leaves
will have a better chance of survival during acclimatization. The control
contained 0.1 f.l.M 2,4-0, 5 f.l.M BAP and 0.25 % activated charcoal. Activated
charcoal is added to the culture medium to reduce necrosis as it adsorbs
toxic components released from tissues. However the presence of activated
charcoal in a culture medium results in the reduction of freely-available
hormones due to the adsorption of hormones by activated charcoal (Ebert
and Taylor, 1990). As the adsorption capacities of different types of charcoal
differ, the level of freely-available hormones present in the medium depends
on the type of the activated charcoal. However, Banaobra III et al. (1998)
reported that any brand of activated charcoal could be used satisfactorily for
in vitro culture of coconut embryos. This indicates that the level of hormones
is not critical in coconut embryo culture . Considering the above facts, and
the results of the present study , it can be concluded that incorporation of the
hormones (2,4-0 and BAP) to the culture medium is not essential for growth
of coconut embryos and a hormone-free medium can be recommended for
coconut embryo culture.
The results of the present study indicated the feasibility of using low
cost ingredients in culture media and using media devoid of hormones for
coconut embryo culture. Thus, the findings will contribute to reduction of
cost in producing coconut plants through embryo culture .
ACKNOWLEDGEMENTS
We would like to thank the staff of the Tissue Culture Division for
their assistance in carrying out the experiments. Thanks are also extended
47
to the Head and the staff of the Biometry Division for their assistance in
statistical analysis of data.
REFERENCES
Areza, M. B. B., Rillo, E. P., Cueto , C. A, Ebert, A W. and Orense, O. D.
(1995). Effects of water quality, pH and state of the medium on growth
and development of coconut embryos in vitro. Cord XI(1) , 1-12.
Assy -Bah , B., Durand-Gasselin, T. and Pannetier, C. (1987). Use of zygotic
embryo culture to collect germplasm of coconut (Cocos nucifera L.).
FAOIIBPGR Plant genetic Resources Newsletter 71, 4-10.
Balaga, H. Y. and De Guzman, E. V. (1971). The growth and development
of coconut 'Makapuno' embryos in vitro. II Increased root incidence and
growth in response to media composition and to sequential culture from
liquid to solid medium . Philippine Agriculture 53(10) ,551-565.
Bonaobra III, Z. S., Rillo, E. P. and Orense, S. D. (1998). Effect of different
brands of activated carbon on growth and development of coconut
(Cocos nucifera L.) embryos in vitro. Cord XIV(2) , 1-11.
Cutter, V. M. and Wilson, K. S. (1954). Effect of coconut endosperm and
other growth stimulants, upon the development in vitro of embryos of
Cocos nucifera. Botanical Gazette 115, 234-240.
De Guzman, E. V. (1970). The growth and development of coconut
'Makapuno' embryos in vitro. I The induction of rooting. Philippine
Agriculture 53(2) , 65-78 .
Ebert, A and Taylor, H. F. (1990). Assessment of the changes of 2,4-
dichlorophenoxyacetic acid concentration in plant tissue culture media in
the presence of activated charcoal. Plant Cell Tissue and Organ Culture
20,165-172.
Eeuwens, C.J . (1978). Effects of organic nutrients and hormones on growth
and development of tissue explants from coconut (Cocos nucifera) and
date (Phoenix dactylifera) palms cultured in vitro. Physio!. Plant 42, 173-
178.
Ganapathi, T. R., Mohan, J. S. S., Suprasanna, P., Bapat, V. A and Rao, P.
S. (1995). A low-cost strategy for in vitro propagation of banana. Current
Science 68(6) ,646-650.
48
JianPing, L., GuangZhi, Z., Gan-Fan, Y., LiPing , P., Luo, J. P., Zheng , G. Z.,
Gan, F. Y. and Peng, L. P. (1996). Cost reduction in cell culture of Panax
ginseng for production of oligosaccharins. Acta Botanica Yunnanica
18(2),139-144.
Karunaratne, S. M. (1987). Tissue culture towards better coconut. Coconut
Bulletin 4(1),6-10.
Karunaratne, S. M. (1988). Short-term in vitro preservation of coconut seed
material: A method to facil itate field collection and transport of coconut
germplasm. Cord IV(2), 40-47.
Karunaratne, S. M., Kurukularachchi, C. and Gamage, C. (1985). A report
on the culture of embryos .of dwarf coconut, Coconut nucifera L. var.
nana in vitro. Cocos 3,1-8.
Karunaratne, S. M., Santha, S. and Kovoor, A. (1991). An in vitro assay for
drought-tolerant coconut germp lasm. Euphytica 53, 25-30 .
Malemnganba, H. Bhattacharya, S. and Deka, P. C. (1994). A new cost
effective embryo culture medium for orchids. Journal of the Orchid
Society oflndia 8(1-2), 67-71.
Okuno, K., Matsumoto, K. and Hisajima, S. (1996). A trial of improvement of
micropropagation of Brassica campestris cv. Kukitachina by cost
simulation. Journal of Society of High Technology in Agriculture 8(4),
253-263.
Orense, J. C. and Pacumbaba, E. P. (1988). Transmission of cadang-
cadang by insects and mites. Annual Report of Philippine Coconut
Authority-Albay Research Center 1988,26-27.
Purnima, H. S. (1997). Manipulation of culture requirements for low cost in
vitro tuberization in potato. Journal of the Indian Potato Association 24(3-
4),128-131.
Rillo, E. P., Paloma, M. B. F., Rodrigues, B. J. and Ignacio, M. T. R. (1988).
Screening of coconut populations for resistance to cadang-cadang
disease using in vitro cultured embryos. Annual Report of Philippine
Coconut Authority-Albay Research Center 1988, 39-40.
Sharma, T. R. and Singh, B. M. (1995) . Simple and cost effective medium
for propagation of ginger (Zingiber ' officinale). Indian Journal of
Agricultural Sciences 65(7),506-508.
49
Sujatha, M. and Chandran, K. (1997). A commercially feasible
micropropagation method for Melia azedarach L. Indian Journal of
Experimental Biology 35(7) ,787-791.
Zheng, G. Z., He, J. B. and Wang , S. L. (1982). Studies on tissue culture of
medicinal plants. VI. Costs in the cell culture of Anisodus acutanggulus.
Acta Botanica Sinica 24(3), 209-215.
50
Table 1: Effect of low cost chemicals on growth of in vitro-cu Itured
coconut zygotic embryos
Treatment Germination Plant Plant height Number of Number of
(%) development (em) leaves per primary
(%) plant roots per
plant
Control 78.9 54.1 9.6 1.4 2.4
Se 83.9 53.0 8.9 1.4 2.3
Kf 79.8 45.0 9.7 1.4 2.2
Se + Kf 79.8 62.8 9.4 1.4 2.1
LSD (p=0.05) NS NS NS NS NS
CV(%) 16.2 20.9 21.7 19.2 21.1
Growth measurements of plants were recorded after four and a half months
of culture. Control = Growth medium with Analar grade chemicals, Sc =
Growth medium with table sugar, Kf = Growth medium with Muriate of
Potash (fertilizer grade KCI), Sc-iKf = Growth medium with table sugar and
Muriate of Potash (fertilizer grade KCI) NS = Not significant.
51
Table 2: Effect of water quality on growth of in vitro-cultured coconut
zygotic embryos
Treatment Germinat ion Plant Plant Number of Number of
(%) . development .height leaves per primary
(%) (em) plant roots per
plant
Control 73.4 75.8 24.8 1.8 2.0
(deion ized
water)
Tap water 73.8 70.4 23.9 1.9 1.8
LSD (P=0.05) NS NS NS NS NS
CV (%) 18.5 28.6 7.7 8.4 9.9
Growth measurements of plants were taken after 6 months of culture.
NS =Not significant.
Table 3: Effect of hormones on growth of in vitro-cultured coconut
zygot ic embryos
Treatment Germ inat ion Plant Plant Number of Number of
(%) development height leaves per primary roots
(%) (em) plant per plant
Control (with 68.2 78.5 26.2 1.9 2.0
0.1 ~M 2,4-0
and 5 11M
SAP)
Medium 81.9 76.6 29.4 2.2 2.0
without
hormones
LSD{p=0.05) NS NS NS 0.2 NS
CV (%) 21.9 8.8 10.4 8.7 11.8
Growth measurements of plants were taken after six months of culture.
NS = Not significant, S = Significant.
52


A cost effective medium for producing embryo-cultured coconut plants

Fernando, M.T.N.

Marikkar, J.M.N.

Jayalath, K.V.N.N.

Subasinghe, S.D.J.N,

COCOS (2000-2002), 14 19-28
Printed ill Sri Lanka
FINANCIAL ANALYSIS OF COCONUT OIL EXTRACTION USING
LOCALLY MANUFACTURED BABY EXPELLERS
M T N Fernando , J M N Marikkar , K V N N Jayalath and S D J N Subas inghe
Coconut Research Institute, Lunuwila, Sri Lanka
ABSTRACT
This paper analyses the financ ial feasib ility of coconut oil extraction using
locally manufactured Baby Expellers, and identifies constraints faced by users
of Baby Expellers. Data were collected by a field survey (February, 1999) of a
purposive sample of fifteen Baby ' Expeller users located in Puttalam ,
Gampaha and Kurunegala districts, using a structured questionnaire,
supplemented with informal discussions. Gross margin (GM), pay-back period
and break-even prices of inputs and outputs were the different economic
indicators employed. The results revealed that oil extraction using Baby
Expellers is econom ically less attractive at the market prices of coconut oil
and copra that prevailed in February 1999. However, there is a marked
improvement of the oil producers' GM when either copra prices are reduced
or oil prices are increased , even by small amounts. The study also revealed
that the profitability of oil extraction is more sensitive to changes of copra
prices than oil prices. As the initial capital requirement to install the Baby
Expeller is substantial , an increased access to loans was recommended to
popularize this technology especially among small and medium
entrepreneurs. The desirability of granting a grace period of over six months
for borrowers was emphasized as the payback period of the initial investment
was also over six months . Low tariffs for palm oil importation and shortage of
copra availability espec ially during lean periods of coconut production appear
to be notable constraints faced by Baby Expeller users. Areas requir ing Rand
D to improve performance of the existing Baby Expeller machines are also
suggested.
INTRODUCTION
Extraction of coconut oil using locally manufactured Baby Expellers ' seems to
be gaining popularity as a cottage industry in many parts of the coconut-
growing areas in the country. The choice of alternative oil-processing
technologies, specially by small and medium size entrepreneurs, is greatly
influenced by the scale of production and the initial capital requirement of
these technologies in addition to the existing resource base of low and
These expellers have an average capacity of abou t 300 kg of coconut oil per 8 hours.
Often these expellers were driven by electric motors.
19
medium-income groups (Anon. 1983, and Etherington and Mahendrarajah,
1997). These attributes of the technology under review seem more
appropriate than the large-scale oil processing technologies, to the existing
resource base of medium-income groups. Recently, the Coconut Research
Institute (CRI) received several inquiries from potential millers who wish to
install Baby Expellers, about the economic feasibility of these mach ines as
well as the likely problems and constraints when using it. However, this
information was not readily available at the CRI, and this motivated the
present study to assess the economic feasibility of oil extraction using Baby
Expellers.
OBJECTIVES
This study aims to:
(i) determine the economic feasibility of oil extraction using locally
manufactured Baby Expellers
(ii) identify the constraints faced by current Baby Expeller users,
and
(Hi) examine the Rand D required to improve the performance of
the locally manufactured Baby Expellers.
METHODS
Data
Data were collected by a field survey, in February 1999, by interviewing a
purposive sample of 15 Baby Expeller users/ located in Puttalam, Kurunegala
and Gampaha districts, through a single visit, using a structured schedule
supplemented with informal discussions.
Analysis
Different economic indicators, namely Gross Margin (GM) per 8 hour working
shift, the Pay-back Period, and the Break-even Price of inputs and outputs
were employed.
2At least, a sample of 30 respondents is required to satisfy the statistical requirements. However,
the population of current holders of locally manufactured Baby Expellers would not exceed
fifty. So, we argue that 15 respondents are a fair representation of the above population. In
terms of mode of operation, size etc., the entire population of baby expellers seems to be
homogeneous. So a purposive sampling was adopted.
20
The GM was computed by deducting total variable cost from gross income.
Pay-back is the length of time required to payoff the initial investment. Break-
even price of inputs means the price of inputs, at which the Baby Expeller
user neither loses nor gains. Similar interpretation applies for break-even
price of output.
The GM for oil extraction using Baby Expellers at prevail ing input and output
prices, and the time taken to pay-back the fixed cost involved in establish ing
the machines were calculated in the base scenario. The sensitivity of the GM
in the base scenario for variat ions in oil and copra prices, each by 10, 20 and
30 per cent was assessed in the sensitivity analys is. In addition , this sect ion
includes the computation of break-even prices of selected inputs and outputs.
The constraints faced by present Baby Expeller users and their suggestions
on R&D needs of the current Baby Expeller were qualitatively assessed
without employ ing any statistical procedures because data were collected
from a limited sample of only 15 respondents.
RESULTS AND DISCUSSION
(I) Base scenario
(i) Gross Margin (GM)
Cost of production (COP) of 1 kg of coconut oil at prevailing prices of inputs
(copra, electricity, fuel wood etc.) was found to be Rs 913 . However, the
current price (as at 26 February, 1999) of 1 kg of coconut oil is Rs 85, which
implies that there is a loss of Rs 6 per kg of oil extracted. Other by-products
such as poonac and pressed oil compensate for this loss. When the incomes
of these by-products are accounted for, the Baby Expeller generates a GM of
some Rs 950 per 8-hour working day (see Table 1).
Table 1:
Item
Cost and returns of coconut oil extraction using baby expellers
Unit Price
OUT PUT (per 8 hrs)
300 kg of oil per 8 hrs
181 kg of poonac per 8 hrs
53.5 kg of pressed oil per 8 hrs
Gross Return
@ Rs 85 per kg
@ Rs 9 per kg
@ Rs 4 000 per barrel of
(200 kg) pressed oil
INCOME (Rs/8
hrs)
25500
1629
1070
28199
COST (Rs/8 hrs)
I US $ = Rs 69.45 as at 26 February 1999.
21
INPUTS
(A) Variable Costs
Labor
2 MD (one man for cutting & drying while
the other man is for work ing
in the expeller)
Materials
- Copra
(535 kg of "kede" copra
is required to extract 300 kg of oil)
- Fuel wood for dryer (1 cubic yard of "wal
dara" is required to obtain 230 kg of oil.
So, to obtain 300 kg of oil, 1.3 cubic yard
of "wal dara"is required .
- Electricity
- Grease
Total variable cost (Rs/8 hrs)
(8) Maintenance Costs (Rs/8 hrs)
Monthly tax to the Ceylon Electricity
Board
Capital depreciation
If the machine were operated 8 hrs a day,
"CONE" has to be welding filled
monthly
If the machine is operated 8 hrs a day,
"WORM" has to be filled and carved out
once in three months
• Filter cloths
Total maintenance cost (Rs/8 hrs)
Total variable costs & maintenance costs
(Rs/8 hrs)
Gross Margin (Rs/8 hrs)
@ Rs 200 per MD
@ Rs 12 400 per candy (or
254 kg) of "kada" cop ra
@ Rs 275 per one cub ic yard of
"waldara"
@ Rs 4.40 per unit of electricity
under industrial category
@ Rs 160 per kg of grease (daily
the machine has to be greased -
about 100g/day)
Unit Price
@ Rs 250 per month (24 working
days a month)
@ Rs 750 per filling (24 working
days a month )
@ Rs 2 000 per filling and carving
out (24 working days a month, and
hence 72 working days per 3
months)
A length (2.5 ft x 2.5 ft) of filter cloth
(@ Rs 3 000) can be used for 1
year (280 working days/ year)
400
26 118
358
275
16
27 167
Cost (Rs) to
extract oil fr om
535 kg of copra
(or cost per 8 hrs
wor king day)
10.40
31.25
28.00
11.00
80.65
27 248
951
Notes : MD - Man-days. 1 US $ = Rs 69.45 as at 26 February, 1999.
Source: Survey of Baby Expellers, 1999.
(ii) Pay-back Period
The fixed costs involved in establishing a Baby Expeller units were computed
(Table 2).
22
Table 2:
Item
Fixed costs involved in installing a Baby Expeller and
accessories
Unit price Rs
i) Baby Expeller machine
10 HP motor (RPM 1440)
• 2 nos. of V-belts
Starter
ii) Cutter
• 5 HP motor
• 2 nos. of V-belts
iii) Dryerb
2 HP motor
Starter
• A V-belt
iv) Filter
• 2 HP motor
A V-belt"
v) Container
(to collect poonac)
vi) Container
(to collect oil)
vii) Tray
(to keep the oil collecting container)
ix) Construction of a concrete base
(to fix the Baby Expeller)
Total
@ Rs 75 OOOa
@ Rs 14500 to 18 000
@ Rs 325/= per belt (Dengil
brand)
(MEM English starter)
(An Indian starter costs
Rs 2 000/=)
@ Rs 3 000 per cutter
@ Rs 12 000 per motor
@ Rs 275 per belt
@ Rs 4 000 per dryer
@ Rs 8 000 per motor
(MEM English starter)
(An Indian starter costs
Rs 2 000/=)
@ Rs 275 per belt
@ Rs 50 000 per filter
75 000
16250
650
3000
3000
12000
550
4 000
8 000
3000
275
50 000
8 000
275
250
250
250
1750
186550
Notes
a-
b -
c -
The price of the Baby Expeller manufactured by the Halmillakele
Engineering works of Wennappuwa was quoted for this examp le.
Use of dryers is not essential if copra is properly dried « 6%
moisture). Alternatively, cut pieces of copra could be sun dried (2 - 3
hours) before feeding into the expeller.
Filter cloths were accounted for under maintenance cost.
As shown in Table 2, the total fixed cost involved in installing a Baby Expeller
is about Rs 187 000. At the prevailing copra and oil prices, 197 eight hour
working days, i.e. over half a year, are required to pay-back the initial
investment cost, which is a substantial blocking of capital for small and
medium scale entrepreneurs. While the economic life is about 40-50 years,
parts like cones, worms etc. have to be completely replaced every 3 -4 years.
(II) Sensitivity analysis
The sensitivity of the GM in the base scenario for changes in oil and copra
prices by plus and minus 10, 20, and 30 per cent was assessed (see Tables 3
and 4 respectively for oil and copra price changes) .
Table 3: GM (Rs/8 hours) of coconut oil extraction at various oil prices
- 30%
- 6699
(- 805%)
- 20%
- 4149
(- 537%)
-10%
- 1599
(- 269%)
Oil price
at
present
price
950
+ 10%
3501
(269%)
+20%
6051
(537%)
+30%
8601
(805%)
Note : Figures in parentheses are the percentage change of GM from the
base scenario.
Source: Table 1.
As shown in Table 2, if the oil price is increased by 10 per cent, i.e. if the
present oil price of Rs 85 per kg is increased to Rs 93.50 per kg, Baby
Expeller users' GM is increased from Rs 950 per 8 hours to Rs 3 501 per 8
hours, representing a 269 per cent increase. The opposite is true for a
. reduction of oil price by 10 per cent. The consequences in response to the
other variations in oil price in Table 3 can be interpreted in the same way.
What all these imply is that the change in oil price by a small amount results
in a great impact on oil producers' GM.
Percentage increase in GM in response to a decrease in copra price by 10
per cent is 275 per cent (Table 4), and how the GM changes in response to
other variations in copra prices (as shown in Table 4) can be explained in the
similar manner. These findings suggest a conclusion as for the case of oil
price changes, i.e. a small change in copra price results in a marked impact
on GM of oil producers.
24
Table 4: GM (Rs/8 hours) of coconut oil extraction at various copra prices
-30% -20%
Copra price
-10% at + 10%
present
price
+20% +30%
4
8788 6175 3561 950 -1662 -4273 -6887
(825%) (550%) (275%) (- 275%) (- 550%) (- 825%)
Note : Figures in parentheses are the percentage change of GM from the
base scenario.
Source: Table 1.
GM increases by some 275 per cent in response to the copra price reduct ion
by 10 per cent, and the corresponding increase in GM in response to the oil
price increase by 10 per cent is 269 per cent (compare Tables 3 and 4). This
implies that the profitability of oil extraction is marginally more sensitive to
-changes in copra prices than that of the oil prices.
• Break-even prices of oil and copra
(a) Break-even price of oil
The break-even price of oil, meaning the price of oil at which the Baby
Expeller holder neither looses nor gains, was found to be Rs 81.80 per kg of
oil" , The current price of oil is Rs 85 per kg, and the break-even price of oil is
just a reduction of some 3.75 per cent less than the current oil price, implying
a fine balance between projects and losses implying that the current oil price
is at a very sensitive position given the higher prices of raw materials, copra in
particular.
(b) Break-even price of copra
Given the current prices of oil, poonac and pressed oil, Rs 85 per kg, Rs 9 per
kg and Rs 20 per kg respectively, the maximum price of copra to break-even
was computed to be Rs 12 -850 per candy (254 kg). This price represents
merely a 2.8 per cent increase from the current copra price of Rs 12 500 per
candy, suggesting that the present price of copra is almost at its maximum
given the prevailing oil price.
Incomes from poonac and pressed oil were also accounted for in calculating the break-
eve n price of oil.
25
(III) .Constraints Faced by Baby Expeller Users
Baby Expeller users pointed out some const raints hindering their business,
and they are:
• Importation of palm oil at concessionary tariffs
The tariff for importation of palm oil has been varying over the time. Orig inally,
the tariff was about 40 per cent , which was then reduced to zero, and
increased up to 8-9 per cent , which is in effect at present (26 February, 1999).
W ith this low tariff rate, palm oil price in the domestic market is as low as Rs
75 per kg. Coconut oil cannot be supplied at such a low price as the COP of
coconut oil is higher (Rs 91 per kg). Therefore, Baby Expeller users were of
the view that the local coconut oil industry has to be protected by imposing
higher tariffs for palm oil imports.
• Shortage of copra in lean crop periods
When highe r prices are offered for coconuts by desiccated coconut (D.C.)
millers, particularly in lean crop periods, the copra manufacturing industry is
seriously affected because coconut growers prefer to sell their cocon uts to
D.C. manufacturers at higher prices without selling to copra manufactures at
low prices. This results in a shortage of copra for the oil industry. A pract ical
example follows.
Some Rs 80 was paid to growers by D.C. millers for nuts equivalent to a kg of
D.C. as at 26 February, 1999.
Approximately 120 kg of D.C. is produced from 1000 nuts). On this D.C
outturn, the price paid by D.C. millers per nut was Rs 9.60. This was a
competitive price and copra manufactures could not pay such a high price to
growers given the prevailing price of coconut oil.
(IV) Research and Development Needs
Baby Expeller users' suggested some areas requmnq research and
development for improving locally manufactured expeller machines".
D.C. out turn is a function of soi l type, pick number, coconut variety, rainfall etc. For
the purpose of this calc ulation, the average D.C. out turn reported by the sample of respondents
(120 kg per 1000 nuts), was used.
6 These sugges tions are howe ver limited to the Baby Expellers manufactu red by the
Halm illakele Industries, We nnappuwa.
26
• Expeller users have found that mixing a little D.C. with copra (some 10-15
kg per candy of copra) imparts an attractive color to coconut oil, and
hence some millers feed copra into expellers mixed with some D.C.
However, the size of the expeller of the present machine is not big enough
to efficiently crush D.C. This was identified by millers as defining requiring
the locally manufactured Baby Expeller which needs to be corrected. The
need for such an improvement of the machine is further emphasized by
the following. Often, villagers bring sliced pieces of sun-dried coconuts
mixed with sun-dried coconut residue to the mills to have the oil extraction
on payments. The efficiency of oil extraction from coconut residue by the
present Baby Expeller is poor, and hence improvement is requ ired for this
reason as well.
• Some millers were of the view that the design of the machine is sound
and the finish is also satisfactory, but the workmansh ip requires
improvement. To quote one example they pointed out , "bearings are of
poor quality and some clamps do not fix tightly".
• As the bars of the bar set are widely spaced, separation of oil from
coconut residues in the expeller is not efficient. Extracted oil is always
mixed with coconut residues. Millers suggested that this defect of the
machine should be overcome.
SUMMARY AND CONCLUSIONS
An attempt was made to assess the economic feasibility of coconut oil
extract ion using Baby Expellers as a cottage industry. A purposive sample of
fifteen entrepreneurs in Puttalam, Kurunegala and Gampaha districts, who
have installed Baby Expeller machines were interviewed in February, 1999 in
a single visit , using a structured questionnaire supplemented with informal
discussions. Such economic indicators as Gross Margin (GM), Pay-back
Period and Break-even Prices of inputs and outputs were employed to
analyze the data. Results revealed that the GM of coconut oil extraction using
Baby Expellers is not attractive under the input and output prices that
preva iled during the period of the survey (February, 1999). However, a small ,
positive change in price of copra and oil resulted in a marked improvement in
terms of returns of the enterp rise. The study also revealed that the profitability
of oil extraction is more sensitive to variations in copra prices than that of the
variations in oil prices, although the differences of changes in GMs in
response to either oil or copra price changes is small. Total fixed cost
required to establish a Baby Expeller is substantial , of the order of Rs 187
000, and hence easier access to industrial loans rr.iqht be an useful means to
encourage especially the small and medium entrepreneurs to invest in these
machines. The pay-back period of the initial capital investment was found to
27
be over a half a year, so granting a grace period of over 6 months for
borrowers might be desirable to allow them adequate time to generate surplus
returns for repaying loans. The local coconut oil industry appears to be
seriously affected by the reduced tariffs for palm oil importation. However,
increased tariff may also lead to efficiency losses and protectionist policies
may not be sustainable in the emerging globalization scenario. So, in depth
stud ies are required on tariffs. As D.C. millers pay competitive prices for
coconuts, particularly in lean crop periods, nut supply to copra manufacturers
is reduced, and this appears to be a serious threat for the sustenance of the
local oil industry. Some useful information on areas requi ring R&D of the
present Baby Expeller machine are also highlighted.
REFERENCES
Etherington, M. and Mahendrarajah, S. (1997) . Economic benefits of direct
micro expelling of coconut oil in the South Pacif ic. Paper presented at the
International Cashew and Coconut Conference. 17-21 February 1997, Dar
es Salaam.
Anon (1983). Small-scale oil extraction from ground nuts and copra.
International Labor Organization Technical Memorandum No.5, 111 pp.
ACKNOWLEDGMENTS
Authors wish to express their gratitude to the Acting Director, CRISL for her
motivation to start this study and providing the approval to publish this paper.
They also wish to thank Mr J L J G Pinto , Former Asst. Information Officer of
CRISL and Mr Rowel Jayathilake, Proprietor of Halmillakele Engineering
works for their useful comments. Assistance of Messrs. Gunarathbanda and
P A B A Caldera in conducting the survey is gratefully acknowledged.
28


Financial analysis of coconut oil extraction using locally manufactured baby expellers

Peiris, T.S.G.

Trends in nut production and utilization

Waidyanatha, U.P.de.S.

'Pol polls: does coconut oil cause artery - clogging?

Samarajeewa, A.D.

Schiere, J.B.

Viets, T.

Ibrahim, M.N.M.

Systems modeling in dairy development under coconut

Ranasinghe, C.S.

Tender king coconut water for the export market

Peiris, T.U.S.

Rajapaksha, S.

»(0 
Prediction of Annual National Coconut Production -
A Stochastic Approach 
T.S.G. PEIRIS 
Coconut Research Institute, Lunuwila, 
Sri Lanka 
T.U.S. PEIRIS 
Coconut Research Institute, Lunuwila, 
Sri Lanka 
and 
S. RAJAPAKSHA 
Coconut Research Institute, Lunuwila, 
Sri Lanka 
ABSTRACT 
Coconut is one of the most economically important crops and it comprises an 
important part of the diet in Sri Lanka. Prediction of annual national coconut 
yield is of vital importance to policy-makers, scientists and coconut growers 
for various reasons. The empirical statistical models based on drought 
indices via monthly rainfall values have failed to predict national yield due to 
many reasons. Properties of the auto correlation of the annual yield (1950-
1988) confirmed that a stochastic model of type ARIMA (2,1,0) was the best 
model to forecast annual national yield. The model does not need any 
external variables. The adequacy of the model was tested based on various 
statistics. The percentage error of the predicted values from 1950 to 1998 
varied from (1-12)%. Based on data 1950-1988, the percentage error for the 
estimated value for the year 1999 is 6.3% with respect to the actual value of 
2830 million nuts. Using data (1950-1999) the estimated values for the years 
2000 and2001 are 2896 and 2705 million nuts. 
Key Words: Stochastic approach, Coconut production, Time series analysis, 
Univariatic model, ARJJMA 
26 
INTRODUCTION 
Coconut is important to Sri Lanka, 
both as a food crop and as an 
industrial crop. The major part of 
the production (70%) is used for 
local consumption and the balance is 
used for export as kernel products 
(24%) and husk nuts (1%). The 
percentage contribution to GDP by 
the coconut sector was 2.6 in 1998, 
which was higher than the 
corresponding percentage by tea 
sector or rubber sector (Anonymous, 
1998). Annual coconut production 
during 1995-1999 fluctuated between 
2540 million nuts (in 1998) and 2830 
million nuts (in 1999) with a mean of 
2654 and standard error of 53 
million nuts. The advanced 
knowledge of coconut production is 
an important tool for policy-makers 
mainly in utilization of nuts for 
various sectors and to search export 
markets. 
In a given location the current crop 
condition, which can be estimated 
from the past weather during the 
growing period of the crop, determines 
its potential yielding ability (Peiris and 
Thattil, 1998). 'fhus the predictive 
models based on climatic variables are 
meaningful in estimating the 
production in a given location. 
However, as the climate variability in 
coconut grown areas in Sri Lanka is 
very complex and not been studied 
deeply, the crop-weather models 
developed to predict national yields 
were not successful in the past (Peiris 
and Mathcs, 1997). The models used 
lo predict national coconut yield are 
discussed by Peiris et al., (1995). 
Presently no statistical methods are 
used to estimate the national coconut 
production in Sri Lanka. 
In tin's paper a univariate time-series 
model based on the yield trend over 
the years was developed to predict the 
annual national yield and compared 
with regression models containing 
climatic variables. 
METHODOLOGY 
Secondary Data 
National annual production data (1950 
to 1999) were collected from the 
annual reports of the Coconut 
Development Authority, Colombo. To 
compute drought indices, monthly 
rainfall data from 1962 to 1997 of 18 
locations in the coconut growing 
areas: Kurunegala, Horombawa, 
Lunuwila, Palugawewa, 
Rajakadaluwa, Ratmalagara, Kirama, 
Mediyawa, Nikaweratiya, 
Polonthalawa, Ridibendi-ela, 
Ratnapura, Gampaha, Galle, Kalutara 
and Matara were obtained from the 
Department of Meteorology, 
Colombo. These stations represent 
four agro-ecological regions of major 
coconut growing regions. The 
27 
locations were selected as the authors 
are presently investigating the long-
term rainfall variability of these 
locations. 
Regression Analysis -Empirical 
Modelling 
Abeywardena (1983) has developed an 
empirical statistical model using data 
from 1963 to 1976 to predict the 
national annual coconut yield a year 
ahead using 'drought indices' of eight 
locations. The locations were 
Tangalla, Kudawewa, Kuliyapitiya, 
Lunuwila, Rajakadaluwa, 
Kurunegala, Negambo and 
Polgahawela. The drought index for a 
given year is derived from the monthly 
total rainfall. The model explained 
95% of yield variation around the 
mean of 1963-1976. The status of 
significance of the model (or the 
status of significance the variables in 
the model) was not mentioned. 
However, as the errors of the 
estimated values for some years were 
large, the model was not used. But no 
alternative climatological models have 
been attempted. 
In our study, the concept of drought 
indices was further tested using more 
recent data. A basic computer 
program was developed to compute 
the drought indices. The drought 
indices were considered as 
independent variables in the regression 
approach. The correlation structure of 
the drought indices among the 
locations does not support strong 
multicollinearity. Of the 18 variables 
of drought indices the best model, 
which is statistically significant and 
explains about 60% of the yield 
variation from 1964 to 1998, is: 
Y = 2996.2937 ( 1 0i)-
1.0201(0.3931)*X1 -
0.8524(o 2945)*X2 + 
0.565 l(o.2993)*X3-
0.7737(o 2993)*X4 + 
0.8471(o.3068)*X5 
(R2 = 0.62; p< 0.001) 
Where, Y = National annual coconut 
production; XI = drought index at 
Lunuwila; X2 = drought index at 
Rajakadaluwa; X3 = drought index at 
Kirama; X4 = drought index at 
Nikaweratiya and X5 = drought index 
at Polonthalawa. The standard errors 
of the parameters are given in 
parenthesis. All the variables in the 
model are significant at the 0.05 level. 
Validity of the Model 
The percentage error for the predicted 
values from 1964 to 1998 for the 
above model varied from 1% to 24% 
(as rainfall data were not available 
prior to 1964 for some stations, yield 
data were used for the period 1964 -
1998). The correlation between the 
observed and predicted values is 
significant (r = 0.79 *",n-36). The 
pattern of the residual was tested and 
28 
found to be distributed randomly. 
These results confirmed the validity of 
the model for the selected data. The 
predicted value for 1999 is 2539 and 
the percentage error is 10.3% as 
against the actual value of 2830. 
Using data from 1964 to 1999, the 
selected five variables of the model 
was tested for three twelve year 
periods: (1964 -1975, 1976 -1987 and 
1988-1999) and it found that only the 
model for 1976 - 1987 was significant. 
Also none of the variables were 
significant in any of the models. 
When the parameters were further 
tested for two 18-year periods (1964 -
1982 and 1983 - 1999) it was found 
that though the models were 
significant, none of the variables were 
significant. It suggests that the model 
(I) is dependent on period of the yield 
data used and thus a unique model 
cannot be recommended irrespective 
of duration of the data. 
Nevertheless the drought indices may 
have biological importance because Y 
is significantly correlated negatively 
with Xj (i = 1,2,3,4,5). It indicates 
that the doughtiness of these places in 
a given year reduces the national yield 
on the following year. Thus the 
concept of drought index should not 
be under estimated. However, it 
suggests the necessity of thorough 
study on long-term rainfall variability 
in many coconuts growing areas. 
Time Series Analysis - Univariate 
Stochastic Approach 
A time series is a set of observation 
generated sequentially in time. The 
annual yield can be considered as a 
discrete time series observed at equal 
time intervals. An intrinsic feature of 
a time series is that the adjacent 
observations are dependent. It was 
found that the auto-correlations of 
yield at the lags 1, 3, and 4 are 
significant al the 5% level. It 
indicates that the yield in a given year 
is correlated to the prior yields and so 
time series approach is appropriate. In 
univariate time series the rainfall 
variability is not directly taken into 
consideration, but it assumed that yield 
fluctuation has occurred due to rainfall 
variability over the years in coconut 
growing areas. 
It is assumed that an observed time 
series Yi, Y2, Yn can be modeled 
by an auto-regressive-integrated 
moving average model of type 
ARIMA(p,d,q) (Box and Jenkins, 
1976) and it can be represented as: 
<|>(B)(I-B)dYt = e(B)e t t= l , 2 , ,n 
Where Y, = national yield in the i"1 
year; n = the number of observations 
for the series; B = the back shift 
operator s.t. B% = Z
 H , <KB) = (1 -
<p,B -<b2B2 d>pBp) and 9(B) = (1 
- 0,B - 9 2B 2 - .... - 0PBP) are two 
polynomials in B; p = the order of the 
29 
auto-regressive part; q = the order of 
the moving average part and d = the 
order of non-seasonal differencing of 
the series. The 's and 9; 's are 
parameters of auto-regressive part and 
moving average part of the model. 
The e; 's are independently and 
identically distributed random 
variables with zero mean and constant 
variance a 2 c. The ARIMA model (2) 
predicts a value in a time series at a 
given time as a linear combination of 
its own past values and past errors. 
Identification of Time Series 
Models 
The auto-correlation function of the 
annual crop production ('original 
series') from 1950 to 1998 indicated 
that the series is non-stationary and 
there is no seasonal or cyclic pattern in 
the national coconut production. The 
series was stationary after a- single 
differencing. The simultaneous 
inspection of the sample ACF and 
PACF indicated that the annual yield 
variation can be represented by the 
models of ARIMA (2,1,0) and 
ARIMA (2,1,1). For each model the 
non-significant chi-square values for 
residuals were observed at lags 
6,12,18, and 24. It indicated that 
residuals were uncorrelated in both 
models and therefore both models 
were considered as possible models to 
represent the annual coconut data. 
Selection of The'Best'Model , 
The maximum likelihood estimates 
of the parameters of ARIMA(2,1,0) 
and ARIMA(2,1,1) models along, 
with their standard errors and 
correlation matrices are given in 
Table 1. 
Table 1 indicates that the ARIMA 
(2.1.0) model has the lowest standard 
errors for parameters and also it has 
the lowest correlation between the 
parameters. The t-ratio. of all the 
parameters in ARIMA (2,1,0) were 
significant at the 5% level. In ARIMA 
(2.1.1) only the second autoregressive 
parameter was significant. It indicates 
that both parameters are significantly 
important for the ARIMA (2,1,0) 
model. Further, the statistics such as 
Akaike's information criterion (AIC), 
and Schwartz' s Bayesian Criterion 
(SBC) were smaller in ARIMA (2,1,0) 
than in ARIMA (2,1,1). Thus it can 
be concluded that the ARIMA (2,1,0) 
model is the most appropriate, to 
model the change in annual coconut 
production. The model is represented 
as: 
Y« = ( l -foB -<b2B2)(1 - B)Y, =e, 
Where <|>i and <t>2 are equal to - 0.26684 
and - 0.66548 respectively. 
30 
Validation of the Predictions 1% to 12% except on two occasions. 
The plot of actual and estimated 
Based on model (3) the percentage values is given in Fig. 1. 
error of the predicted values for the 
period from 1950 to 1998 varied from 
Table 1: Estimates of the model parameters using maximum likelihood method. 
Model Parameters Std. error of 
the 
parameters 
Correlation matrix 
ARIMA(2,1,0) <J>, = - 0.26684 
<|>2 = - 0.66548 
0.10701 
0.10557 
1.000 
0.157 1.000 
ARIMA(2,1,1) G,= 0.07177 
<h = - 0.22699 
d>2 = - 0.65982 
0.22043 
0.16167 
0.11039 
1.000 
0.738 1.000 
0.224 0.255 1.000 
Figure J: Plot of actual and estimated national annual coconut production. 
31 
The correlation between the actual and 
estimated values is significant (r = 
0.87***,n=50). The estimated value 
using model (3) for the year 1999 is 
2650 million nuts as against the actual 
value of 2830. The percentage error 
is 6.3%, compared to 10.3% error with 
the regression model. 
When two 30-year periods: 1950 -
1979 and 1968 -1997 were considered 
it was found that the best fitted model 
is ARIMA (2,1,0). By including the 
most recent forecast value, prediction 
was done for one year ahead using 
ARIMA (2,1,0) type model. The 
estimated yield for the period 2000 to 
2004 are 2896, 2705 , 2765, 2875 and 
2815 million nuts, respectively. 
Time Series Analysis - Bivariate 
Stochastic Approach 
In addition to the pattern of climate 
('rainfall')) in the previous year, other 
factors that could influence the 
national yield are the nut price of the 
previous years and the amount of 
fertilizer used in the previous years. 
It was found that the national coconut 
production has no significant cross 
correlation with the nut prices or the 
amount of annual fertilizer used in 
the previous years. Because of this 
reason Abeywaredena (1983) 
excluded the fertilizer as a dependant 
variable for his national production 
function. Therefore, for bi-variate 
time series those two factors need not 
to be considered. 
DISCUSSION 
The need for a national yield 
forecasting model has been felt by the 
coconut industry since the failure of 
Abeywardena's model during 1980 si • 
Therefore the model developed in this 
paper is of immense value.for various 
activities of the industry. As the rate 
of national consumption for culinary 
use is 90 nuts per head per year, the 
number of nuts required for culinary 
purposes can be estimated based on 
the rate of annual population increase 
(1%). Therefore the balance nuts after 
culinary use can be computed using 
predicted values for future years. The 
advance knowledge of balance nuts 
will be useful in planning to share nuts 
among various activities in the 
coconut industry such as production of 
desiccated coconut, coconut oil, copra, 
coconut milk/cream/milk powder and 
fresh nut exports. Based on the,past 
trend of nut utilization for the above 
activities, it can be predicted that there 
will not be a scarcity of coconut in 
future for the coconut industry. 
The ARIMA class of time series 
models is a powerful tool for 
forecasting national coconut 
production than regression models. 
The advantage of time series model is 
that prediction of yield can be done 
even more than one year ahead 
32 
without any external variables. Using 
the predicted values for 2000 and 
2001. the predicted national coconut 
production for the years 2002 to 2004 
arc 2765, 2875 and 2815 million nuts 
Time scries data are often 
contaminated with outliers due to the 
influence of unusual and non-
repetitive events. Thus the 
incorporation of the method suggested 
by Chung and Liu (1993) for the time 
series packages will be beneficial to 
tackle such data. 
CONCLUSION 
A univariate stochastic model of type 
ARIMA (2,1,0) can be used to predict 
the annual coconut production in Sri 
Lanka for efficient planning of the 
coconut industry. 
ACKNO WLEDG M ENT 
Part of the work was funded by the 
National Science Foundation, 
Colombo, Sri Lanka. 
REFERENCES 
Abeywardena, V. (1983). Effect of 
moisture stress and irrigation on 
yield of coconut in Sri Lanka. (In) 
Coconut Research and 
Development. N.M. Nayar (Ed.), 
pp. 98-106. Wiley Eastern 
Limited, New Delhi. 
Anonymous (1998). Annual Report 
of the Central Bank of Sri Lanka. 
Colombo, Sri Lanka. 
Box, G.E.P. and G.M. Jenkins (1976). 
Time Series Analysis: Forecasting 
and Control. San Francisco, 
Holdcn-Day. 
Chung, C. and L-M. Liu (1993). 
Forecasting time series with 
outliers. Journal of Forecasting, 
12: 13-35. 
Peiris, T.S.G., R.O. Thattil and R. 
Mahindapala (1995). An analysis 
of the effect of climate and 
weather on coconut (Cocos 
nuciferu). Experimental 
Agriculture 31:(4) 451-460. 
Peiris, T.S.G. and D.T. Mathes 
(1997). The analysis of long term 
rainfall variability at selected 
locations in Sri Lanka. 
Proceedings of the International 
Congress on Modelling and 
Simulation 06-11 December, 1997 
Hobart, Australia. 69-74. 
Peiris, T.S.G. and R.O. Thattil 
(1998). The study of climate 
effects on the nut yield of coconut 
using parsimonious models. 
Experimental Agriculture 34: 
189-206. 


Prediction of annual national coconut production-a stochastic approach

The role of coconut industry in Sri Lanka towards the world trend on kernel products.

Mewan, K.M.

Liyanage, A.C.

Everard, J.M.D.T.

Random amplified polymorphic DNA (RAPD) markers reveal geneticrelatedness of tea (Camellia sinensis L..) clones in Sri Lanka

Gunathilake, H.A.J.

Arambewewela, L.

Ratnayake, H.

Rajapakse, S.

cocos (2000-2002), 14, 87-96
Printed in Sri Lanka
FEASIBILITY OF GROWING MEDICINAL PLANTS IN COCONUT LANDS
OF THE WET ZONE OF SRI LANKA
H A J Gunathilake, L Arambewewela*, H Ratnayake and 8 Rajapakse
Coconut Research Institute, Lunuwile , Sri Lanka
ABSTRACT
There is a growing demand for plant materials for use in perfumery and
ayurvedic preparations. Although much research has been carried out on
intercropping coconut lands, little attention has been given for medicinal
plants.
Ten selected medicinal plants were tested at Walpita Estate in the Low
Country Wet Zone of Sri Lanka. Yield and chemical quality of the species
grown under coconut and in the open field were assessed.
Piper longum under coconut showed higher plant yield and piperine content
than when grown in the open field. Kaempheria galanga and Plumbago
indica also showed the same trend. Yield difference of Adhathoda vasica
and Aloe vera under coconut and in open field was not significant. The other
species, Plectranthus spp, Solanum xanthocarpum, Hibiscus abelmoschus,
Withania somnifera and Cassia angustifolia did not perform well in AER
WL3 and require further evaluation in other agro-c1imatic regions.
INTRODUCTION
The Coconut Research Institute has been engaged in research on
intercropping in coconut lands, with both annual and perennial crops such
as coffee , pepper, pineapple, ginger, etc., and has a considerable database
and exper ience in this field (Gunathilake and Liyanage, 1996). However,
lack of experience and knowledge on intercropping of medicinal plants in
coconut lands is a serious drawback in popularising such crops.
8ri Lanka imports approximately Rs. 80-100 million worth of medicinal plant
materials for use in perfumery and ayurvedic preparations. A major
component of annual plant species such as Solanum xanthocarpum (8:
Katuwelbatu), Cassia angustifolia (S: Senehekola), Withania somnifera (S:
Ashawagandha), Piper longum (S: Tippili), Kaempferia galanga (S:
Ingurupiyali) , Hibiscus abelmoschus (S: Kapukinissa) etc. The chemical
• Natural Product Division, ITI, Bauddhal oka Mawatha , Colombo 7
87
activity of some of the imported dried plant material is less than the desired
level due to long periods of storage and transportation before they are used.
The objectives of this work were to study the feasibility of growing several,
hither to imported medicinal plants species under coconut and to assess the
chemical quality of the produce cultivated under these conditions.
MATERIALS AND METHODS
This study was carried out at Walpita Estate, Kotadeniyawa in the Agro
Ecological Reagion WL3 where annual rainfall is approximately 2000 - 2200
mm. The age of the Coconut plantation was 46 years and available sunlight
was approximately 60% under coconut, which was planted at 8.5 m x 8.5 m
spacing (150 palms/ha). The soil was well drained and moderately deep
Boralu series . (Sornasiri et al., 1994).
Selected medicinal plants listed in Table 1, were established in 4.0 m x 6.0
m plots prepared in the center of the coconut square. The plots were
arranged in a Complete Randomized Block Design with three replicates per
species. At planting , 500 g of concentrated Super Phosphate was applied
to each plot followed by the application of 10 kg of dried cattle manure every
three months during the experimental period.
A similar experiment was conducted in an open area (without coconut) for
comparision. .
Growth and reproductive parameters were recorded monthly and samples
were taken for assessment of yield and chemical quality at the end of the
first and second years, after planting. Agrochemicals were not used for pest
and disease control.
RESULTS
Yield
The yields obtained over the first year, both in the open and under coconut
are presented in Table 2. Plectranthus spp (Iriweriya) failed to produce any
yield due to high mortality. Iriweriya plants underboth conditions (open and
under coconut) were susceptible to collar rot disease caused by Pythium
spp and succumbed after two months. Under coconut Piper longum (Tippili)
produced two and half times the yield obtained in the open field.
Kaempheria galanga (Ingurupiyali) also yielded more under coconut than in
the open. The yield of the other species tested was less under coconut
than in the open field .
. 88
Extrapolated yields for one ha of land in the open (10,000 m2) and under
coconut (excluding 2m radius for coconut manure circle - 6,600 m2) are
given in Table 3. It showed that Piper longum has yielded more under
coconut although the effective area is 2/3 under coconut. Kaempheria
galanga, Aloe vera, Plumbago indica, Adhathoda vasica and Withania
somnifera continued to produce a yield in the second year also (Table 4).
Except for Withania and Plumbago, the yields under coconut of the other
three species during the second year were higher than during the first year
(Table 4). Extrapolated fresh leaf yield of Aloe vera in the second year was
12.6m.t.lcoconut ha (Table 5).
Chemical Quality
A chemical analysis of the medicinal plants/parts are presented in Table 6.
Piper longum, Kaempheria galanga and Phumbago indica had higher
concentrations of the active chemical ingradient in plants grown under
coconut compared to those grown in the open. In contrast, Solanum
xanthocarpum, Withania somnifera and Cassia angustifolia produced a
lower concentration of the active chemicals under coconut than in open
sunlight.
Comparing yield and volatile oil content of Kaempheria in the first and
second years, those two characters behaved inversely i.e. in the second
year , yield was higher but volatile oil was lower (Table 6). While Aloe vera
and Plumbago produced higher contents of chemicals in the second year
than in the first year.
DISCUSSION
Piper longum and Kaempheria galanga performed well under coconut
showing their shade loving nature. These two species probably originated
under shade in their natural habitats. The other species produced low yields
under coconut possibly due to competition above or below the ground level.
As the supply of nutrients and water was abundant, it can be assumed that
the yields were reduced due to insufficient sunlight. This was clearly evident
with Solanum xanthocarpum, Hibiscus abelmoschus, Withania somnifera
and Cassia angustifolia. Although the yield per plant of Aloe vera, Plumbago
indica and Adhathoda vasica under coconut was slightly reduced, their
performance under coconut could be considered satisfactory. Aloe vera,
Plumbago, Adhathoda, Kaempheria and Withania survived and carried over
in to the second year after planting, as 'they are perennials. Aloe vera,
Adhathoda and Kaempheria yielded fairly well in the second year. Plimbago
could be maintained up to the end of the second year its yield was reduced
considerably.
89
In regard to both chemical quality and yield, Piper longum and Kaempheria
performed well under coconut.
Yield and chemical content of Plumbeqo indica grown under coconut
behaved inversely, with the yield reduction being compensated by the better
quality of extracts showing that this plant is also adaptable to the coconut
situation.
Assessed in terms of yield and quality Solanum xanthocarpum, Hibiscus
abelmoschus, Withania somnfera and Cassia angaustifolia failed to tolerate
conditions under coconut in WL3. Plectranthus spp needs further
investigation on its susceptibility to soil-borne diseases.
CONCLUSION
Piper longum, Kaempheria galanga and Plumbago indica are clearly
medicinal plants with a good potential to be developed as intercrops under
coconut in the wet zone due to their shade loving nature. Adhathoda vasica
and Aloe vera also show promise. The other species tested should be
experimented under different agro-c1imatic conditions for their adaptability.
ACKNOWLEDGEMENTS
Financial support given by the Council for Agricultural Research Policy
(CARP) is qratefully acknowledged. We are also qrateful to Mr. R M S
Rathnayake for initiating the programme, Mr. M A Dayawansa and Mr. R
Thambugala are also acknowledged for their help in conducting field trials
and chemical analyses.
REFERENCES
Gunathilake, H A J and Liyanage, M de S (1996). Multiple Cropping under
Coconut (In) Agroforestry for Sustainable Development in Sri Lanka.
(Eds. P A Huxley, D M S H K Ranasinghe)
Somasiri, L L W Nadarajah, N Amarasinghe, Land Gunathilake, H A J
(1994). Land suitability Assessment of Coconut Growing Areas in the
Coconut Triangle. Occasional Publication Series No.3, Coconut
Research Institute of Sri Lanka, December, 1994.
90
Ta
bl
e
1:
Th
e
sp
ac
in
g
an
d
de
ns
ity
o
fw
hi
ch
th
e
m
ed
ic
in
al
pl
an
ts
w
er
e
te
st
ed
Sc
ie
nt
ifi
c
N
am
e
Co
m
m
on
N
am
e
Sp
ac
in
g
(cm
)
D
en
si
ty
Pl
an
ts
/h
a
Pl
an
ts
/c
oc
on
ut
ha
*
Pi
pe
rl
on
gu
m
Ti
pp
ili
75
x7
5
17
,5
00
11
,7
00
Ka
em
ph
er
ia
ga
la
ng
a
In
gu
ru
pi
ya
li
60
x4
5
36
,0
00
24
,0
00
Al
oe
ve
ra
Ko
m
ar
ic
a
50
x3
0
66
,0
00
44
,0
00
Pl
um
ba
go
in
di
ca
R
at
hn
ito
l
60
x6
0
27
,0
00
18
,
00
0
\0
Pl
ec
tra
nt
hu
s
sp
p
Iri
w
er
iy
a
30
x3
0
11
1,
00
0
72
,0
00
So
la
nu
m
Ka
tu
w
el
ba
tu
45
x1
5
14
2,
50
0
95
,0
00
x
a
n
th
oc
ap
um
H
ib
is
cu
s
Ka
pu
ki
ni
ss
a
60
x6
0
27
,
00
0
18
,0
00
a
be
lm
os
ch
um
W
ith
an
ia
so
m
n
ife
ra
Am
uk
ka
ra
60
x3
0
54
,0
00
36
,
00
0
Ca
ss
ia
a
n
gu
st
ifo
lia
Se
ne
he
ko
la
60
x3
0
54
,0
00
36
,0
00
Ad
ha
th
od
a
va
si
ca
Ad
ha
th
od
a
60
x9
0
18
,0
00
12
,0
00
Ar
ea
e
xc
lu
di
ng
2.
0
m
ra
di
us
fo
rc
o
co
n
u
tm
a
n
u
re
ci
rc
le
(ef
fec
tiv
e
la
nd
a
re
a
fo
ri
nt
er
cr
op
pi
ng
in
o
n
e
he
ct
ar
e
o
fc
o
co
n
u
ti
s
6,
60
0
m
2 )
Ta
bl
e
2:
Yi
el
d
in
gr
am
pe
rp
la
nt
o
ve
r
th
e
fir
st
ye
ar
in
th
e
o
pe
n
an
d
u
n
de
rc
o
co
n
u
t
M
ed
ic
in
al
Pl
an
t
M
ed
ic
in
al
ly
u
se
fu
lp
ar
t
Yi
el
d
at
1
ye
ar
af
te
r
pl
an
tin
q
(g
/pl
an
t)
O
pe
n
Un
de
rc
o
co
n
u
t
Th
ip
pi
li
Be
rri
es
(dr
ied
)
13
.4
a
32
.4
b
+
24
1.
8
In
gu
ru
pi
ya
li
R
hy
zo
m
e
(dr
ied
)
39
.2
'a
45
.2
b
+
15
.3
Ko
m
ar
ic
a
,
Le
av
es
(fr
es
h)
17
19
a
11
45
.0
b
-
33
.4
-
.
0
R
at
hn
ito
l
R
oo
ts
(dr
ied
)
13
1.
9
a
86
.9
b
-
34
.1
N
Ad
ha
to
da
To
ta
lp
la
nt
(dr
ied
)
18
0.
0
a
13
2.
5
b
-
26
.4
Ka
tu
w
el
ba
tu
*
To
ta
lp
la
nt
(dr
ied
)
24
.8
a
10
.1
b
-
59
.3
Ka
pu
ki
ni
ss
a
Se
ed
s
(dr
ied
)
27
.
0
a
5.
1
b
-
81
.1
Am
uk
ka
ra
R
oo
ts
(dr
ied
)
48
.8
a
15
.2
b
-
68
.9
Se
ne
he
ko
la
Le
av
es
(dr
ied
)
4.
2
a
1.
3
b
-
69
.0
Iri
w
er
iya
Sh
oo
ts
(dr
ied
)
n
il
n
il
*
H
ar
ve
st
ed
a
t5
Y2
m
o
n
th
s
a
fte
rp
la
nt
in
g
(W
ith
in
co
lu
m
ns
,v
a
lu
es
sh
ar
in
g
a
co
m
m
o
n
le
tte
rd
o
n
o
td
iff
er
si
gn
ific
an
tly
a
tP
=0
.0
5)
Ta
bl
e
3:
Ex
tr
ap
ol
at
ed
yi
el
d
o
fd
iff
er
en
tm
ed
ic
in
al
pl
an
ts
in
th
e
o
pe
n
an
d
u
n
de
rc
o
c
o
n
u
t
M
ed
ic
in
al
M
ed
ic
in
al
ly
Yi
el
d
at
1
ye
ar
af
te
r
%
di
ffe
re
nc
e
pl
an
ts
u
se
fu
lp
ar
t
pl
an
tin
g
c
o
m
pa
re
d
to
o
pe
n
O
pe
n
Un
de
r
c
o
c
o
n
u
t
kg
/h
a
kg
/c
oc
on
ut
ha
*
Th
ip
pi
li
Be
rri
es
(dr
ied
)
23
5
a
37
9
b
+
61
.1
\0
In
gu
ru
pi
ya
li
R
hy
zo
m
e
(dr
ied
)
1,
41
0a
1,
08
4
b
-
22
.
5
'
'
'
'
)
Ko
m
ar
ic
a
Le
av
es
(fr
es
h)
.
11
3,
45
4
a
50
3,
80
0
b
-
55
.6
R
at
hn
ito
l
R
oo
ts
(dr
ied
)
3,
56
1
a
1,
56
4
b
-
56
.1
Ad
ha
to
da
To
ta
lp
la
nt
(dr
ie
d)
3,
24
0
a
1,
59
0
b
-
50
.9
Ka
tu
w
el
ba
tu
To
ta
lp
la
nt
(dr
ied
)
3,
53
4
a
96
0
b
-
72
.8
Ka
pu
kl
ni
ss
a
Se
ed
s
(dr
ied
72
9
a
92
b
-
87
.4
Am
uk
ka
ra
R
oo
ts
(dr
ied
)
2,
63
5
a
54
7
b
-
79
.
2
Se
ne
he
ko
la
Le
av
es
(dr
ied
)
22
7
a
47
b
-
79
.4
*
Ar
ea
a
va
ila
bl
e
a
fte
re
xc
lu
di
ng
2.
0
m
ra
di
us
fo
rc
o
co
n
u
tm
a
n
u
re
ci
rc
le
(W
ith
in
co
lu
m
n
,
va
lu
es
sh
ar
in
g
a
co
m
m
o
n
le
tte
rd
o
n
o
td
iff
er
si
gn
ifi
ca
nt
ly
a
tP
=0
.0
5)
Ta
bl
e
4:
Yi
el
d
in
gr
am
pe
rp
la
nt
in
th
e
se
co
n
d
ye
ar
in
th
e
o
pe
n
an
d
u
n
de
r
c
o
c
o
n
u
t
M
ed
ic
in
al
M
ed
ic
in
al
ly
Yi
el
d
(g
/pl
an
t)
%
di
ffe
re
nc
e
pl
an
ts
u
se
fu
lp
ar
t
c
o
m
pa
re
d
to
o
pe
n
O
pe
n
Un
de
r
IQ
c
o
c
o
n
u
t
.
;:..
In
gu
ru
pi
ya
li
R
hy
zo
m
e
(dr
ied
)
62
.7
a
76
.6
a
+
30
.8
Ko
m
ar
ic
a
Le
av
es
(fr
es
h)
2,
31
2.
8
a
2,
87
6.
7
a
+
24
.4
R
at
hn
ito
l
R
oo
ts
(dr
ied
)
56
.7
a
59
.5
a
-
4.
9
Ad
ha
to
da
To
ta
lp
la
nt
(dr
ied
)
1,
32
7.
0
a
1,
07
2.
2
b
-
19
.2
Am
uk
ka
ra
R
oo
ts
(dr
ied
)
15
.8
a
6.
5
b
-
58
.9
(W
ith
in
co
lu
m
n,
va
lu
es
sh
ar
in
g
a
co
m
m
o
n
le
tte
rd
o
n
o
td
iff
er
si
gn
ifi
ca
nt
ly
a
tP
=O
,0
5)
Ta
bl
e
5:
Ex
tr
ap
ol
at
ed
yi
el
d
o
fs
ev
er
al
m
ed
ic
in
al
pl
an
ts
in
th
e
ye
ar
af
te
r
pl
an
tin
g
in
th
e
o
pe
n
an
d
u
n
de
r
c
o
c
o
n
u
t
M
ed
ic
in
al
pl
an
ts
M
ed
ic
in
al
ly
u
se
fu
lp
ar
t
Yi
el
d
O
pe
n
kg
/h
a
Un
de
r
c
o
c
o
n
u
t
kg
/c
oc
on
ut
ha
*
\0
In
gu
ru
pi
ya
li
R
hy
zo
m
e
(dr
ied
)
2,
25
4
a
1,
83
9
b
V
I
Ko
m
ar
ic
a
Le
av
es
(fr
es
h)
15
2,
64
4
a
12
6,
57
3
b
R
at
hn
ito
l
H
oo
ts
(dr
ied
)
1,
53
1
a
1,
07
1
b
Ad
ha
to
da
To
ta
lp
la
nt
(dr
ied
)
23
,8
86
a
12
,8
86
b
Am
uk
ka
ra
R
oo
ts
(dr
ied
)
85
3
a
23
4
b
*
Ar
ea
a
va
ila
bl
e,
e
xc
lu
di
ng
a
fte
r2
.0
m
ra
di
us
fo
rc
o
co
n
u
tm
a
n
u
re
ci
rc
le
(W
ith
in
co
lu
m
n
,
va
lu
es
sh
ar
in
g
a
co
m
m
o
n
le
tte
rd
o
n
o
td
iff
er
si
gn
ifi
ca
nt
ly
a
tP
=0
.
05
)
Ta
bl
e
6:
Ch
em
ic
al
s/
in
gr
ed
ie
nt
s
an
al
ys
is
o
fm
ed
ic
in
al
pl
an
ts
M
ed
ic
in
al
M
ed
ic
in
al
ly
Ch
em
ic
al
s/
Pe
rc
en
ta
ge
pl
an
ts
im
po
rt
an
t
in
gr
ed
ie
nt
s
O
pe
n
Un
de
r
Co
co
nu
t
pa
rts
a
s
s
o
c
ia
te
d
1'
YA
P
2
YA
P
1
YA
P
2Y
A
P
Th
ip
pi
li
Be
rri
es
Pi
pe
rin
e
0.
64
-
0.
82
(d
rie
d)
In
gu
ru
pi
ya
li
R
hy
zo
m
e
Vo
ta
til
e
o
il
7.
41
3.
25
8.
73
3.
31
(dr
ied
)
Ko
m
ar
ic
a
Le
av
e
s
Fr
e
sh
pl
an
t
2.
35
6.
30
2.
23
6.
40
(fr
es
h)
jui
ce
R
at
hn
ito
l
R
oo
ts
To
ta
le
xt
ra
ct
14
.7
28
.8
6
22
.9
7
27
.
60
\0
(d
rie
d)
-
e
0
\
Ka
tu
w
el
ba
tu
W
ho
le
pl
an
t
A
lk
al
oi
ds
in
*
1.
9
*
0.
81
.
-
(dr
ied
)
"
ro
o
ts
Ka
pu
ki
ni
ss
a
Se
ed
s
-
n
.
a
.
n
.a
.
n
.
a
.
n
.
a
.
(dr
ie~
.
Am
uk
ka
ra
R
oo
ts
Al
ka
lo
id
s
0.
91
0.
26
0.
26
0.
27
(dr
ied
)
St
er
ei
ds
Se
ne
he
ko
la
Le
av
es
Se
nn
os
id
e
1.
50
-
1.
20
Ad
ha
to
da
W
ho
le
pl
an
t
Es
se
nt
ia
le
il
n
.a
.
n
.
a
.
n
.
a
.
n
.a
.
(dr
ied
)
*
Sa
m
pl
es
co
lle
ct
ed
a
t5
Y2
a
n
d
15
m
o
n
th
o
fa
ge
fo
r
1
YA
P
a
n
d
2
YA
P
re
sp
ec
tiv
el
y
n
.a
-
n
o
ta
n
a
lys
ed
YA
P
-
ye
a
rs
a
fte
r
pl
a
n
tin
g


Feasibility of growing medicinal plants in coconut lands of the wet zone of Sri Lanka

National Science Foundation of Sri Lanka

Not a member yet