PERTAfllHfl A Journal of Biological, Physical and Social Sciences

Vol. 12 No. 1 April, 198^

Contents
Section 1 : Biological Sciences
An Electrophoretic Study of Natural Populations of the Cocoa Pod Borer, Conopomorpha cramerella (Snellen)
from Malaysia. - Rita Muhammad, S.G. Tan Y.Y. Gan, 5. Rita. S. Kanasar and K.Asuan. 1
A Rust Disease on Gendarussa vulgaris Nees. Caused by Puccinia thwaitesii Berk. - Kamaruzaman Sijam and
T.K Lim. 7
A Rot of Detached Durian Fruits Caused by Sclerotium rotfsii. - T.K Lim and Kamnruzaman Sijam. 11
Nutrient Composition of Selected Cooked and Processed Snack Foods. -E.S. Tee, S. SiH Mizura, A. Anuar,
Ft Kuladevan, S.I Young, S.C. Khor and S.fC Chin 15
Dietary Self-Selection by Laying Hens Offered Choices of Feed. - Ramlah Hamid, R.I. Hutagalung and P.N. Vohra 27
Determination of Meat Content in Processed Meats Using Currently Available Methods. - AS. Babji, P.M. Ooi
and A.Abdullah. 33

Communications
Kesan Pendedahan Racun Karbari! terhadap Tumbesaran dan Pembiakan Aphis cracrivora Koch. - Yusoj bin
Ibrahim and Wong Kuan Yeow. 43
The Rate of Litter Production in Mangrove Forest at Siar Beach, Lundu, Sarawak. - Saberi Othman. 47
Diseases in Dogs: Necropsy Observation. - KF. Chooi, T. Pandiyaraja and A.R. Sheikh-Omar. 53

Section II : Physical Sciences

Chemical Characterisation of Podzols in Sabah, East Malaysia. - Marcus Jopony and Tan Ying Chiu. 59
Changes in Surface Potential of Activated Carbon Due to Adsorption of Ions. - M. Badri and Karen A. Grouse. 65
A Simple and Clean Method for Methoxymethylation of Phenols. -Faujan B.H. Ahmad and J. Malcolm Bruce. 71
The Differences in Thermal Breakdown of H2PtCl6 and cis-Pt(NH3)2 d2 in Carbon Furnace Atomic Absorption.
- AM. Jelas Haron and Wan Md. Zin Wan Yunus. 79
On the Higher Order Edge-Connectivity of Complete Multipartite Graphs. - Y.H. Peng, C.C. Chen and KM. Koh. 83
On Robust Alternatives to the Maximum Likelihood Estimators of a Linear Functional Relationship.
- Mokhtar B. Abdullah. 89
Satu Kajian tentang Getaran Terusik ke atas Selaput Segi Empat Sama dengan Menggunakan Teori Usikan.
- Zainul Abidin Hassan dan Salwa bt. Abu Bakar. 99
Communications
Lime Requirements of Highly Weathered Malaysian S o i l s . - / Shamshuddin. 107
Effects of Effective Size in Rapid Sand Filtration. - Ahmad Jusoh, Megat Johari Megat Mohd Moor and Abdul
Halim Ghazali. 113

Section III : Social Sciences

Estimating the Income and Substitution Effects on the Demand for Poultry Meat. - ZainalAbidin Mohamed
and Roslan A. Ghajfar. 119
Community Leader Participation in the Extension Programme Decision-Making Process. - Aminah Ahmad
and iMverne B. Forest. 125

A scientific Journal published by ISSN 0126-6128

UNIVERSITI PERTANTAN MALAYSIA
 ARCHIVE COPY
(Please Do Not Remove)

EDITORIAL BOARD

CHIN HOONG FONG (Chief Editor)

EDDIE CHIEW FOOK CHONG (Business Manager)

SULAIMAN MOHD YASSIN ASIAH BTE MOHD ZAIN

ANG KOK fEE TAN HOCK SENG
KWOK CHEE YAN KAREN ANNE GROUSE BADRI
MAT YUSOFF ABDULLAH KAMIS AWANG
SUMANGAIA PILIAI (Secretary)

PERTANIKA is a scientific journal published thrice a year (April, August and December) by Universiti
Pertanian Malaysia (University of Agriculture, Malaysia) in which papers in Bahasa Malaysia and English in
any area aligned with the work done at the faculties of the University appear. Currently these include
Agriculture, Forestry, Veterinary and Animal Science, Food Science and Technology, Resource Economics
and Agribusiness, Engineering, Fisheries and Marine Science, Science and Environmental Studies, Extension
and Continuing Education, Education and Social Studies, Human Development and Consumer Studies.

PERTANIKA welcomes original reports in English or Bahasa Malaysia of research not previously or simul-
taneously published in any scientific or technical journal from the staff of Universiti Pertanian Malaysia and
other local and overseas institutions and organisations. Contributions are reviewed by a panel of consultants
whose names appear in the last issue of each volume.

PERTANIKA is currently abstracted by the following: Agrindex, Biological Abstracts, Chemical Abstract,
Nutrition Abstracts, Animal Breeding Abstracts, Field Crop Abstracts, Forestry Abstracts, Forest Production
Abstracts, Herbal Abstracts, Horticultural Abstracts, Indian Veterinarian, Plant Breeding Abstracts, Reviews
in Applied Entomology, Reviews in Plant Pathology, Soils and Fertilizers and Veterinary Bulletin.

Article in triplicate should be submitted to the Chief Editor, PERTANIKA, Universiti Pertanian Malaysia,
Serdang, Selangor, Malaysia.

Subscription Rates(per year)

Malaysia/Singapore Overseas

Individual $45,000 US$30.00

Institutions $70.00 US$35.00

Overseas subscribes please add US$6.00 per issue for airmail surcharge.

Cheques/Bank drafts should be made payable to UNIVERSITI PERTANIAN MALAYSIA and sent to,
PERTANIKA, Universiti Pertanian Malaysia, 43400 UPM Serdang, Selangor Darul Ehsan, Malaysia.
 PERTflniHfl
A Journal of Biological, Physical and Soc^f¥cienc)fes M 0 H D ZAMRI SAAD
J & J
Deputy Director
XT u i A -l iftoo PuDlication Division
N u m b e r 1 , A p r i l 1989 Research M a n a g e m e m Centre
Universit* Putra Malaysia
Contents Serdang, Seiangor Darul Ehsan
Section 1 : Biological Sciences
An Electrophoretic Study of Natural Populations of the Cocoa Pod Borer, Conopomorpha cramerelta (Snellen)
from Malaysia. - Rita Muhammad, S. G. Tan K. Y. Gan, S. Rita, S. Kanasar and KAsuan. 1
A Rust Disease on Gendarussa vulgaris Nees. Caused by Puccinia thwaitesii Berk. - Kamaruzaman Sijam and
T.K Urn. 7
A Rot of Detached Durian Fruits Caused by Sclerotium rolfsii. - TK Lim and Kamaruzaman Sijam. 11
Nutrient Composition of Selected Cooked and Processed Snack Foods. - E.S. Tee, S. Siti Mizura, A. Anuar,
R Kuladevan, SJ Young, S.C. Khar and S.K Chin 15
Dietary Self-Selection by Laying Hens Offered Choices of Feed. - Ramlah Hamid, R.I. Hutagalung and P.N. Vohra 27
Determination of Meat Content in Processed Meats Using Currently Available Methods. - A. S. Babji, P.H. Ooi
and A.Abdullah. 33

Communications
Kesan Pendedahan Racun Karbaiil terhadap Tumbesaran dan Pembiakan Aphis cracdvora Koch. - Yusof bin
Ibrahim and Wong Kuan Yeow. 43
The Rate of Litter Production in Mangrove Forest at Siar Beach, Lundu, Sarawak. - Saberi Othman. 47
Diseases in Dogs: Necropsy Observation. - K.F. Chooi, T. Pandiyaraja and 4-ft Sheikh-Omar. 53

Section II : Physical Sciences

Chemical Characterisation of Podzols in Sabah, East Malaysia. — Marcus Jopony and Tan Ying Chiu. 59
Changes in Surface Potential of Activated Carbon Due to Adsorption of Ions. - M. Badri and Karen A. Grouse. 65
A Simple and Clean Method for Methoxymethylation of Phenols. - Faujan BM, Ahmad and J. Malcolm Bruce. 71
The Differences in Thermal Breakdown of H2PtC!6 and cis-Pt(NH s ) 2 CI<2 in Carbon Furnace Atomic Absorption.
- Md. Jelas Haron and Wan Md. Zin Wan Yunus. 79
On the Higher Order Edge-Connectivity of Complete Multipartite Graphs. - Y.H. Peng, C.C. Chen and KM. Koh. 83
On Robust Alternatives to the Maximum Likelihood Estimators of a Linear Functional Relationship.
- Mokhtar B. Abdullah. 89
Satu Kajian tentang Getaran Terusik ke atas Selaput Segi Empat Sama dengan Menggunakan Teori Usikan.
- Zainul Abidin Hassan dan Salwa bt. Abu Bakar. 99
Communications
Lime Requirements of Highly Weathered Malaysian Soils.-/. Shamshuddin. 107
Effects of Effective Size in Rapid Sand Filtration. - Ahmad Jusoh, Megat Johari Megat Mohd Noor and Abdul
Halim Ghazalz. 113

Section III : Social Sciences

Estimating the Income and Substitution Effects on the Demand for Poultry Meat. - ZainalAbidin Mohamed
and Roslan A. Ghaffar. 119
Community Leader Participation in the Extension Programme Decision-Making Process. - Aminah Ahmad
and iMverne B, Forest. 125
 Section I
Biological Sciences

.
Pertanika 12(1), 1-6(1989)

An Electrophoretic Study of Natural Populations

of the Cocoa Pod Borer, Conopomorpha cramerella
(Snellen) from Malaysia.
RITA MUHAMAD, S.G. TAN, Y.Y GAN, S. RITA,
S. KANASAR and *L ASUAN.
Departments of Plant Protection, Biology, and Biotechnology
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia.

Key words: Conopomorpha cramerella; cocoa pod borers; rambutan fruit borers; electromorphs;
polymorphisms.

ARSTRAK
Pengorek buah koko dari Tawau, Sabah dan Sua Betong, Negeri Sembilan dan pengorek buah rambutan dari
Serdang dan Puchong, Selangor dan Kuala Kangsar, Perak, Malaysia telah dianalisa secara elektroforesis
dalam usaha untuk mendapatkan diagnosis elektromorf antar kedua biotip Conopomorpha cramerella.
30 enzim dan protein-protein umum telah dapat ditunjukkan pada zimogram-zimogram tetapi tidak ada satu
pun yang boleh digunakan sebagai penanda diagnosis antara pengorek buah koko dengan pengorek buah
rambutan. Frekwensi alil-alil untuk 8 enzim polimorf juga dipaparkan.

ABSTRACT
Cocoa pod borers from Tawau, Sabah and Sua Belong, Negeri Sembilan and rambutan fruit borers from
Serdang and Puchong, Selangor and Kuala Kangsar, Perak, Malaysia were subjected to electrophoretic analysis
in an effort to find diagnostic electromorphs between these two biotypes o/Conopomorpha cramerella. Thirty
enzymes and general proteins were successfully demonstrated on zymograms but none of them could serve as
diagnostic markers between cocoa pod borers and rambutan fruit borers. The allelic frequencies for 8 polymorphic
enzymes are presented.

INTRODUCTION bromae cocoaL.) and that which attacks rambutan

The cocoa pod borer, Conopomorpha cramerella
(Snellen) (Lepidoptera: Gracilariidae) is a
major cocoa pest in Sabah State, Malaysia but
until late 1986 it was only present as a minor
pest of rambutan in Peninsular Malaysia (Loke
etal 1986, Ling et at 1987). Both the cocpa
pod borer and the rambutan fruit borer are
known scientifically as C. cramerella. An outbreak
of the cocoa biotype occurred for the first time
in the states of Malacca and Negeri Sembilan
in Peninsular Malaysia in September and
November 1986 respectively. Fortunately the
cocoa pod borer had not been detected
elsewhere in the Peninsula (Chin 1987).
This outbreak enabled us to collect both the
biotype that attacks the pods of cocoa (Theo-
(Nephelium lappaceum L.) fruits from the
Peninsula within a period of three months
(November 1986 to January 1987) although
unfortunately not from the same locality. We
were also able to obtain the cocoa biotype from
Tawau, Sabah in January 1987. With these
samples, we attempted to find diagnostic
electormorphs between the two biotypes
of C.cramerella.
MATERIALS AND METHODS
The following samples were available for electro-
phoretic analysis: cocoa pod borers from Sua
Betong Estate (SB) near Port Dickson in Negeri
Sembilan collected in November 1986 and from
the Tawau district of Sabah collected in January
 RITA MUHAMAD, S.G. TAN, Y.Y. GAN, S. RITA, S. KANASAR AND K. ASUAN
1987 (Tl); rambutan fruit borers from Ladang
7 in the campus of Universiti Pertanian Malay-
sia, Serdang, Selangor collected in January 1987
(LI); from Kuala Kangsar, Perak State (KK)
collected in January 1987 and from Puchong,
Selangor state collected in January 1987 (PI)
and in September 1987 (P2). The adult insects
were frozen at - 70°C until they were used
for electrophoresis.
Sample homogenization and polyacry-
lamide gel electrophoresis were done as in
Rusnah et al (1985). Various buffer systems
such as TEB (Green 1977), CA-7 (Steiner and
Joslyn 1979), TEMM (Spencer et al 1964)
and that of Varvio-Aho et al (1980) were used
in our screening for diagnostic electromorphs
between cocoa pod borers and rambutan fruit
borers. The staining procedures used were from
Steiner and Joslyn (1979), Harris and
Hopkinson (1976), Menken (1980), Shaw
and Prasad (1970) and Munstermann (1979).
Isoelectric focusing was performed as in Tan
et al (1982) using LKB ampholytes with pH
ranges of 5-7 and 4-6 after which general protein
were stained for by using Coomassie Brilliant
Blue R250.
The polymorphic markers, phospho-
glucomutase (PGM), esterase (EST) (Halmy
et al. 1987), a-glycerophosphate dehydro-
genase (ct-GPDH), peptidase (PEP) (Rusnah
et al 1985) and malate dehydrogenase were
typed on CA-7 buffer with 0.1% Kodak Photo-
flo incorporated into the gel while hexokinase
(HK), fluorescent esterase (FE) and malic
enzyme (ME) were typed on the Varvio-Aho
buffer (Tan et al 1987). PEP was also typed on
the Varvio-Aho buffer.
RESULTS AND DISCUSSION
At least two adult insects each of cocoa pod
borers from Tawau (Tl) and Sua Betong (SB)
and at least two adult insects each of rambutan
borers from Ladang 7 (LI) and Puchong
(PI or P2) were screened for the presence of
diagnostic electromorphs of the following
enzymes that were successfully demonstrated
on zymograms: aldehyde dehydrogenase,
aldehyde oxidase, trehalase, glucose dehydro-
genase, aconitase, alcohol dehydrogenase,
fumarase, phosphoglucose isomerase, xanthine
dehydrogenase, isocitrate dehydrogenase,
PERTANIKA VOL. 12 NO. 1, 1989
glucose-6-phosphate dehydrogenase, 6-phospho-
gluconate dehydrogenase, diaphorase,
fluorescent acid phosphatase, fucose de-
hydrogenase, L-threonine dehydrogenase,
succinic dehydrogenase, uridine mono-
phosphate kinase, glutamate oxaloacetate trans-
minase, fructose 1, 6 diphosphatase, pyruvate
kinase and acid phosphatase.
However, no consistent diagnostic elec-
tromorphs were observed between rambutan
fruit borers and cocoa pod borer for any of the
above enzymes. No diagnostic protein bands
were observed between cocoa pod borers and
rambutan fruit borers although numerous bands
were seen on the gel for each insect sample
after isoelectric focusing runs on pH ranges
4-6 and 5-7. The allelic frequencies for the
polymorphic systems are presented in Table 1.
None of these polymorphic systems showed any
diagnostic electromorphs between CPB and RB.
The data for populations T2 (CPB from Tawau,
collected in December 1984) and P3 (RB from
Puchong, collected in January 1985) are from
Rusnah et al (1985) and that for L2 (RB from
Ladang 7 UPM collected in March-April 1985)
is from Halmy et al (1987). The data for HK,
MDH, FE-2 and ME-1 for populations Tl, SB,
P2 and LI are obtained from Tan et al (1987)
which reported in detail the occurrence of
polymorphisms for these enzymes in C. cramer-
ella, their biochemical characteristics and their
proposed modes of genetic control. All the
allelic frequencies from these papers are
compiled in Table 1 together with data reported
here for the first time for comparative purposes
and in order to calculate their heterozygosity
values. Data for EST-1 is not presented here as
it could not be reliably typed in this study
because the quality of the EST-1 bands deterio-
rated on storage during the time span of about
ten months from the time the samples were
first collected to the time the electrophoretic
analysis was finally completed. Fortunately, the
bands for the other polymorphic systems
remained consistent and could be typed reliably
during the study period.
The addition of Kodak Photoflo to the
CA-7 gel improved the resolution of the PGM
bands so that another allele, PGM9g which was
not detected previously (Halmy et al 1987)
could now be typed. PEP was initially typed on
 TABLE 1
Gene frequency data for Esterase-2 (EST-2), phosphoglucomutase (PGM), peptidase-2 (PEP-2), a-glycerophosphate dehydrogenase (a-GPDH)
hexokinase (HK), malate dehydrogenase (MDH), fluorescent esterase -2 (FE-2) and malic enzyme (ME-1) in natural populations of
Conopomorpha cramerella (Snellen) from five localities in Malaysia.

ROPH ORETH
Allelic frequencies

Locus &
Allele 77 12 LI L2 PI P2 P3 KK
c;
PGM
105 0.231 ± 0.029 NA 0.344 ± 0.033 0.352 ± 0.038 0.302 ± 0.032 0.340 ± 0.067 0.500 ± 0.083 NA 0.304 ± 0.068
± 0.074 w
100 0.639 ± 0.033 0.543 ± 0.034 0.512 ± 0.039 0.476 ± 0.034 0.420 ± 0.070 0.361 ± 0.080 0.522
98 _ 0.080 ± 0.038 0 0.044 ± 0.031 0
;• 0.038 ±0.013 0.033 ± 0.012 0.049 ± 0.017 *n
97 0.087 ± 0.020 0.080 ± 0.019 0.086 ± 0.022 0.221 ± 0.028 0.160 ±0.052 0.140 ± 0.058 0.130 ± 0.050 Z
N 104* . ^ •
106* 81* 106* 25* 18* 23* £
z H 0.538 r 0.591 0.593 0.642 0.640 0.611 0.565
• i . ' '• •
EST-2
o 102 0.279 ± 0.031 NA 0.361 ± 0.031 0.375 ± 0.039 0.355 ± 0.037 0.406 ± 0.087 0.375 0.070 NA 0.417 ± 0.082
r 100 : 0.553 ± 0.035 0.574 ± 0.032 0.487 ± 0.041 0.428 ± 0.038 0.469 ± 0.088 0.479 ± 0.072 0.389 ± 0.081 Sc
97 • 0.168 ± 0.026 -,• r • 0.066 ± 0.016 0.138 ± 0.028 0.217 ± 0.032 0,125 ± 0.054 0.146 ± 0.051 0.194 ± 0.067
N 104* v
';" •-• K 122* 76* 83* 16* 24* 18*
p
is
H
PEP-2 (CA-7 buffer)
0.567 , " - • '.

0.525 0.592 0.614 - 0.563 0.708 0.444

1
c
F 0.947 ± 0.036 0.919 ± 0.022 1.000 1.000 r :• ^ 1.000 NA ^ 1.000 1.000
s NA
S
N
H
0.053 ± 0.036
19*
0.105
0.081 ± 0.022
74
0.162
! • * & • • • •
0
18*
0
0

0
19*
:.
:. ;

-. °
••
'
1 8 *
0
r
u °
41*
0
0
14*
0
1
0
HK
104
1.00 1.000
0 . NA ;^ 0
1.000
0.111
0.889
± 0.025 2
± 0.025
NA, 0.042 ± 0.041
0.958 ±0.041
0
1.000
NA 0.080 ± 0.038
0.920 ± 0.038
1
N
H
102
0
.-. ' r
.'•• 112
0
81
0.173 ••,•:
12*
0.083
28
0 •••
25*
0.080
i
PEP'2-(Varia-Aho's buffer)
102 0.144 ± 0.037 *NA ^ 0 0 NA NA 0 NA NA
1
^ !i • - '

100 0.800 ± 0.042 ».932 ± 0.027 0.895 ± 0.050 0.940 ± 0.034

— 98 0.056 t 0,024 0.068 ± 0.027 0.105 ± 0.050 0.060 ± 0.034
N 45* 44* 19* 25*
H 0.311 . 0.136 0.211 0.12
a-cmn
F= 100 0.995 ± 0.005 1.000 0.996 ± 0.004 0.956 ± 0.019 NA 1.000 0.981 ± 0.019 0.989 ± 0.005 0.917 ±0.046
S = 97 0.005 ± 0.005 0 0.004 ± 0.004 0.018 ±0.012 0 0.019 ±0.019 0.011 ±0.005 0.083 ± 0.046
104 0 0 0 0.026 ±0.015 0 0 0 0
N 98* 86 118* 57* 18* 26* 184 18*
0.088 2
H 0.010 0 0.008 0 0.038 0.022 0.167 X
MDH
104 0.012 ± 0.008 NA 0.044 ±0.014 0.024 ± 0.013 NA 0 0 NA 0
100 0.988 ± 0.008 0.956 ± 0.014 0.976 ± 0.013 1.000 1.000 1.000 CO
N 84 103 63 5* 24 20*
0.024 0.087 0.048 0 0
p
PI
H 0
FE-2
105 0.125 ± 0.034 NA 0.241 ± 0.040 0.159 ± 0.055 NA NA 0.233 ± 0.055 NA NA
IKA\

100 0.729 ± 0.045 0.518 ± 0.047 0.659 ± 0.072 0.433 ± 0.064 O

97 0.021 ± 0.015 0.027 ±0.015 0 0.100 ± 0.039
0 95 0.125 ± 0.034 0.205 ± 0.038 0.182 ± 0.058 0.217 ± 0.053
90 0 0.009 ± 0.009 0 0.017 ± 0.017
N 48 56 22 30 2
0 H 0.438 0.554 0.455 0.633
ME-1
19*

102 0 NA 0.514 ±0.035 0.346 ± 0.054 NA NA 0.250 ± 0.060 NA 0.304 ± 0.068 2:

100 0.471 ± 0.038 0.486 ± 0.035 0.615 ± 0.055 0.750 ± 0.060 0.696 + 0.068
98 0.529 ± 0.038 0 0.038 ± 0.022 0 0
N 85 104 39 26 23* >
II 0.447 0.471 0.462 0.346 0.435 a
H 0.280 0.075 0.285 0.322 0.630 0.277 0.311 0.0178 0.255
III 0.240 0.244 0.321 0.342 0.242 0.250 G
Tl - Tawau 1987, T2 • Tawau 1984, SB = Sua Betong 1986, LI = Ladang 7 1987, L2 = Ladang_7 1985, PI = Puchong Jan., 87, P2 = Puchong Sept., 87, P3 = Puchong
1985, KK = Kuala Kangsar 1987, N = sample size, H= Heterozygosity, H - Mean Heterozygosity, HI * Mean Heterozygosity based on PGM, EST-2, HK, ct-GPDH & MDH
only. NA = Not analysed. * • Data presented for the first time in this paper.
 AN ELECTROPHORETIC STUDY OF NATURAL POPULATIONS OF CONOPOMORPHA CRAMERELLA
the CA-7 buffer system used by Rusnah et al
(1985) but it was subsequently typed on the
Varvio-Aho buffer because on this buffer
three alleles could be typed for PEP-2 where-
as on the CA-7 buffer, only two alleles could
be recognised.
PGM, EST-2, PEP-2 (Varvio-Aho buffer),
FE-2 and ME-1 were polymorphic in all the
populations that were analysed for these
enzymes. Slight differences were observed in
the allelic frequencies between the various geo-
graphical populations and between populations
from the same locality but collected at different
times. However, there were no clear cut differ-
ences between the allelic frequencies and
heterozygosities for populations of rambutan
fruit borers and cocoa pod borers.
PEP-2 on CA-7 buffer showed poly-
morphism only in the Tawau sample of cocoa
pod borers where alleles PEP-2* and PEP-2S
(Rusnah et al. 1985) were present whereas in all
the other populations typed only allele PEP?
was present. On the Varvio-Aho buffer, allele
PEP-2102 was only present in the Tawau popu-
lation whereas alleles PEP-2 10° and PEP-298 were
present in the Tawau and Sua Betong cocoa
pod borer populations and in the Ladang 7 and
Puchong populations of rambutan fruit borers.
Hexokinase was only polymorphic in the
Ladang 7, Puchong and Kuala Kangsar popu-
lations of rambutan fruit borers in which both
alleles HKm4 and HK100 were present but the
most common allele in these populations,
namely HKm, was fixed in the cocoa pod borer
populations of Tawau and Sua Betong. a-GPDH
was polymorphic in the Tawau, Sua Betong,
Ladang 7, Puchong and Kuala Kangsar
populations with alleles a-GPDH 10° (a-GPDH¥
in the terminology of Rusnah et al. 1985) and
a-GPDH97 (or -GPDHS) being present but allele
a -GPDH104 was only present at a low frequency
in the Ladang 7 population. MDH was only
polymorphic in the cocoa pod borer popula-
tions of Tawau and Sua Betong and in the
rambutan fruit borer population of Ladang 7
where allele MDH104 occurred at low frequen-
cies. The Puchong and Kuala Kangsar popula-
tions were fixed for the common MDH 10° allele.
As regards ME-1, ME-m was the commoner allele
in the cocoa pod borer populations of Tawau,
ME-1 102 in the cocoa pod borer population of
PERTANIKA VOL. 12 NO. 1, 1989
Sua Betong while the most frequent allele in
the rambutan fruit borer populations from
Ladang 7, Puchong and Kuala Kangsar was
ME-110°.
Hence, while differences do exist in allelic
frequencies between populations of C.cramerella
from different geographical areas of Malaysia,
such differences also exist between populations
from the same locality but collected at different
times. Unique alleles were present in certain
populations but the commoner alleles were
present in all populations. We have therefore
been unable to find any diagnostic gene or
allele or even to build up a biochemical key
that would enable us to distinguish electro-
phoretically between cocoa pod borers and
rambutan fruit borers in our present study
using electrophoresis.
H, the average heterozygosity value,
calculated based on data for PGM, EST-2, HK,
a-GPDH and MDH which are available for Tl,
SB, LI, PI, P2 and KK showed values in the
range of 0.240 to 0.342 with no consistent
differences between the two populations of
cocoa pod borers and the four populations of
rambutan fruit borers. (PI 8c P2 differed in
their times of collection from the same locality).
Rusnah et al. (1985) had suggested the
possibility that a large scale electrophoretic
survey of C, cramerella from Sabah and Peninsular
Malaysia may reveal the existence of two species.
However, in this study we were unable to find
any evidence to support this suggestion. Fur-
ther work using more powerful biochemical
genetic tools like restriction fragment length
polymorphisms (RFLP) should be done to
determine whether diagnostic biochemical
markers really do not exist between cocoa pod
borers and rambutan fruit borers. However,
financial constraints have prevented us from
using these techniques.
ACKNOWLEDGEMENT
This work was supported by research grant no.
50281: Genetics from Universiti Pertanian
Malaysia. We thank Dr Tay Eong Bok, State
Director of Agriculture, Sabah for providing
the Tawau Samples, Dr. Loke Wei Hong
(MARDI) for providing the Kuala Kangsar
samples, the Federal Department of Agri-
culture, Malaysia and Mr. Liau Siau Suan from
 RITA MUHAMAD, S.G. TAN, Y.Y. GAN, S. RITA, S. KANASAR AND K. ASUAN

Chemara Research Centre for their invaluable tera, Yponomeutidae) Speciation, (Studies in Yponomeu-
cooperation and assistance during the col- tidae). PhD. dissertation. University of Leiden, The
lection of the Sua Betong samples. Netherlands.
MUNSTERMANN, L.E. 1979. Isozymes ofAedes aegypti.
REFERENCES Phenotypes Linkage and Use of Genetic Analysis of
Sympatric Species population in East Africa. PhD.
CHIN, F. 1987. Views on the Eradication Pro- dissertation. University of Notre Dame, U.S.A.
gramme for the Cocoa Pod Borer in Peninsular
RUSNAH MOHD. SAAHLAN, RITA MUHAMMAD, S.G.
Malaysia. In Proceedings of the Symposium on
TAN, Y.Y. GAN, HALMY AHMAD. 1985. Biological
Management of the Cocoa Pod Borer, pp. 176-180.
Polymorphisms of Peptidase a n d a - G P D H in
Kuala Lumpur.
Natural Populations of the Cocoa Pod Borer
GREEN, C.A. 1977. A Sex-limited Esterase in the Conopomorpha cramereUa (Snellen). Journal of Plant
Accessory Glands of Males of Anopheles funestus. Protection in the Tropics. 2: 49-52.
Mosquito News 3T\ 46-48.
SHAW, C.R. and R. PRASAD. 1970. Starch Gel
HALMY, AM S.G. TAN, M. RITA, and Y.Y. GAN. 1987. Electrophoresis of Enzymes. A Compilation of
Esterase and Phosphoglucomutase Polymorphism Recipes. Biochemical Genetics 4: 297-320.
in the Rambutan Fruit Borer, Conopomorpha
SPENCER, N., D.A. HOPKINSON and H.
cramereUa (Snellen). Journal of Plant Protection in
PARRISH. 1964. Phosphoglucomutase Poly-
the Tropics. 4: 101-105.
morphism in Man. Nature (Lond.) 204: 742-745.
HARRIS, H., and D.A. HOPKINSON. 1976. Handbook
STEINER, W.W.M. and DJ. JOSLYN. 1979. Elec-
of Enzyme Electrophoresis in Human Genetics.
North - Holland. Amsterdam. trophoretic Techniques for the Genetic Study of
Mosquitoes. Mosquito News. 39: 35-54.
LING, A.H., C.C. YEW, and T.S. KOH. 1987. Ex-
TAN, S.G., Y.Y. GAN, and K. ASUAN. 1982. Transfer-
perience in Handling the Cocoa Pod Borer
Outbreak in Peninsular Malaysia. In Proceedings of rin Subtyping in Malaysians and Indonesians from
the Symposium on Management of the Cocoa Pod Borer,
North Sumatra. Human Genetics 60: 369-370.
pp. 129-137. Kuala Lumpur. TAN, S.G., M. RITA, YY. GAN, and S. RITA. 1988.
LORE, W.H. O. MOHD. SHAMSUDIN, and A.K.
Hexokinase, Malate Dehydrogenase, Fluorescent
ZAM. 1986. Occurrence and Status of Acrocer- Esterase and Malic Enzyme in the Cocoa Pod
Borer, Conopomorpha cramereUa Snellen. Pertanika
cops cramereUa Snellen in Peninsular Malaysia. In
Cocoa and Coconuts: Progress and Outlook. Kuala
11(1): 7-13.
Lumpur: ed E. Pushparajah and P.S. Chew, pp VARVIOAHO, SIRKKA-LJISA, and P. PAMILO. 1980. A
265-267. Incorporated Society of Planters, New Buffer System with Wide Applicability. Isozyme
Bulletin. 13: 14.
MENKEN, S.BJ. 1980. AUozyme Polymorphism and the
Speciation Process in Small Ermine Moths (Lepidop (Received 9 November, 1988)

6 PERTANIKA VOL. 12 NO. 1, 1989

Pertanika 12(1), 7-10(1989)

A Rust Disease on Gendarussa vulgaris Nees. Caused by Puccinia

thzvaitesii Berk.
KAMARUZAMAN SIJAM and T.K. LIM
Department of Plant Protection,
Faculty of Agriculture,
Universiti Pertanian Malaysia,
43400 UPM Serdang, Selangor D.E. Malaysia

Keywords: Leaf rust disease; medicinal plant; Puccinia thivaitesii.

ABSTRAK
Gendarussa vulgaris Nees., suatu tanaman herba perubatan yang biasa ditemui, telah didapati teruk
dijangkiti oleh penyakit karat Ciri-ciri/pertanda penyakit ini dapat dilihat pada lesion-lesion cekung atau
cembung yang berwarna kelabu-ungu, dikelilingi oleh halo kuning yang bias di alas permukaan adaksial
daun dan lesion-lesion cembung atau cekung yang sepadan dipermukaan abaksial yang bertentangan. Daun-
daun yang berpenyakit gugur sebelum matang. Agen penyebab telah dikenalpasti sebagai Puccinia thwaitesii
Berk. Kulat ini mengeluarkan kedua-dua uredospora dan teliospora di atas pustul yang sama dan adalah
merupakan karat autesius mikrosiklik (hemi-form).

ABSTRACT
Gendarussa vulgaris Nees., a common medicinal herb, was found to be severely infected by a rust disease.The
disease was characterised by well defined, concave or convex, purplish-gray, discoloured lesions surrounded by
a wide yellow halo on the adaxial surface and corresponding convex or concave lesions on the opposite abaxial
surface. Infected leaves abort prematurely. The causal agent was identified as Puccinia thwaitesii Berk. The
fungus produced both uredospores and teliospores on the same pustule and is a microcyclic (hemi-form),
autoecious rust.

INTRODUCTION pathogen. Some aspects of the epidemiology of

Gendarussa vulgaris Nees (syn. Justicia genda-
russa L.), a monotypic genus of the family
Acanthaceae, and colloquially known as 'ganda
rusa' ('ghanda' in Sanskrit meaning fragrance
and 'rusa' in Malay meaning deer) is a common
medicinal herb and is also used as an
ornamental hedge plant. The plant is renown
as a traditional cure for many ailments, such as
stomach swelling, lunacy, snake-bite, rheuma-
tism, debility, and as a decoction for worms
(Burkill 1966). It is also used as a protective
charm to ward off evil by superstitious folks.
The crop plant in the Universiti Pertanian
Malaysia (UPM) herbarium garden was recently
found to be severely infected by a leaf rust
disease. This paper reports on the symptomato-
logy, identification and characterisation of the
the rust are also discussed.
MATERIALS AND METHODS
Rus tinfected leaves of G. vulgaris were collected
from the medicinal plant herbarium in UPM
campus for microscopic examination and
histopathological investigations. Histopatho-
logical sections of the infected leaves with
varying degrees of rust pustule development
were made following standard histological
methods of Sass (1958). Adjacent plants and
undergrowth which included weeds near
Gendarussa host plants were also examined
thoroughly for rust incidence. Fallen diseased
leaves and plant host debris in the immediate
vicinity were also collected for study.
 KAMARUZAMAN SIJAM AND TX LIM
RESULTS AND DISCUSSION
Symptomatology
The rust causes well-defined, discoloured, pur-
plish-gray, necrotic lesions, 6-15 mm in diame-
ter, surrounded by a distinct yellow halo on the
upper leaf surface (Fig. 7). Such lesions may be
convex (bulged) or concave (sunken) on the
upper leaf surface and correspondingly con-
cave or convex on the opposite leaf surface.
Fig 1. Upper leaf surface showing the distinct concave or
convex, discoloured, necrotic lesions and yellow halo.
Severely diseased leaves often become
distorted because of such lesions. When adja-
cent lesions coalesce, the yellow peripheral area
enlarges. Subsequently the whole leaf turns
yellow with discrete pockets of purplish-gray
islands (Fig, 2). One to 16 lesions can occur on
any one leaf. Infected leaves abort prematurely
in great numbers.
Fig. 2. Gendarussa plants severely infected by Puccinia
thwaitesii.
8 PERTANIKA VOL. 12 NO. 1, 1989
Telia are most frequently observed on the
lesions. They appear as minute, pin-size, pow-
dery, brownish-black specks which are formed
in well defined clusters on the undersurface of
the lesions surrounded by a light green halo on
the abaxial leaf surface (Fig. 3). On some le-
sions, uredia can be seen surrounding the telia
forming tiny light-brown blisters usually in a
concentric ring or rings around the cluster of
telia on the same lesion or sometimes scattered
amongst the telia. Both telia and uredia are
frequently found to be overgrown by a Fusarium
mycoparasite which imparts a whitish, cottony
growth over the purplish-brown rust pustules
(Fig. 3). The parasitism of this mycoparasite will
be discussed in a separate paper.
Fig. 3. Lower leaf surface showing the purplish-brown rust
pustules, some of which are parasitized by a Fusa-
rium mycoparasite which produced a white turf of
mycelia over the pustules.
Transverse sections of the infected leaves
reveal that the hyphae of the fungus proliferate
between the cells of the mesophyll and palisade.
The rust pustules protrudes through the lower
leaf surface through stomatal openings.
Pathogen
The rust fungus is identified as Puccinia thwai-
tesii Berk. Two stages of the rust fungus can
be observed on some necrotic lesions on the
leaves viz. teliospores (Fig 4) and urediospores
(Fig. 5). However, teliospores are more fre-
quently and abundantly produced. Teliospores
are dark-brown, two-celled, ellipsoidal, or
clavate, slightly constricted at the septum and
slightly attenuated below, 36.8 ± 4.8 x 19.7 ± 1.7
|im. The teliospore wall is smooth and very
 A RUST DISEASE ON GKNDARUSSA VULGJLRIS NEES. CAUSED BY PUCCINIA rHWAITESII BERK.
Fig. 4. Micrograph of a telia showing the thick-walled,
two-celled teliospores. Note the two nuclei (arrowed
a) and the pore (arrowed b). Bar = 18 \Lm.
finely reticulate of approximately 2.6 ± 0.2 ^m
thick at the side and 3.65 ±1.1 (Xm thick above,
with one or more pores; pedicels are basal and
persistent, paler than the spores, smooth, with
variable length of 74.2 ± 17.0 (im. Teliospores
always give rise to basidiospores (Laundon
1967). However, this was not observed on the
host lesions. Uredia are amphigenous, occur-
ring in concentric rings in association with telia
on some lesions. Urediospores are pale-brown,
ellipsoidal to obovoidal, unicellular, and
echinulate, 21.2 ± 2.0 x 24.4 ± 2.5 mm, usually
2 pores at the equatorial region. The uredospore
wall is amber, and 1.5 |Xm thick. Uredospores
are borne singly on pedicels which arise in
compact clusters.
This rust was first reported by Parandekar
and Ajrekar (1932) in India on the same host.
They observed that this rust fungus produced
only teliospores which germinated immediately.
This was also reported by Laundon (1963).
Payak (1952) reported that the diploidisation
process in the telial primodia of Pucdnia thwai-
tesii occurred through cell fusion rather than by
nuclear migration. However, in rusts, telia can
originate from aeciospores or uredospores
(Laundon, 1967).
Epidemiology
The rust fungus has a microcyclic life cycle
(Hawksworth et al 1983), producing ure-
dospores and teliospores which will produce
basidiospores. No pycnial or aecial stages were
present. Using the terminology of Peterson
PERTANIKA VOL. 12 NO. 1, 1989
Fig. 5. Micrograph of a uredia showing the echinulated
thin-walled uredospores borne on closely-packed
pedicels. Bar = 42 \im.
(1974), this rust can be regarded as hemi-form
as it possesses stage II (uredospores), stage III
(teliospores) and stage IV (basidiospores). Since
only G. vulgaris was infected and not the weed
species in the immediate vicinity, the fungus
appears to be autoecious without an alternate
host. Teliospores were more abundantly pro-
duced than urediospores. Most of the diseased
leaves which fell to the ground contained more
teliospores than urediospores indicating telio-
spores to be more important in disease spread.
Fallen infected leaves appeared to serve as an
important inoculum source. Disease spread
within plants was observed to be by water splash
and crawling insects, but between plants, wind
may be more important.
Many of the telia and uredia were found
to be naturally parasitized by a mycoparasite
which formed whitish tufts of mycelia over the
rust pustules. The Fusarium mycoparasite may
play an important role in the secondary spread
of the rust pathogen. However, further studies
are needed to confirm the identity of the
mycoparasite and its relationship and impor-
tance to the epidemiology of the rust disease.
REFERENCES
BURKILL, I.H. 1966. A Dictionary of the Economic
Products of the Malay Peninsula. Vol. 1. Ministry of
Agriculture and Cooperatives, Kuala Lumpur,
Malaysia.
HAWKSWORTH, D.L., B. C. SUTTON and G.C.
AlNSWORTH. 1983. Dictionary of the Fungi,
9
 KAMARUZAMAN SIJAM AND T.K. LIM

Seventh Edition. Commonwealth Mycological
Institute, Kew, Surrey.
LAUNDON, G.F. 1963. Rust Fungi I: On Acan-
thaceae. Mycobgical Papers no. 89. Commonwealth
Mycological Institute, Kew, Surrey, England.
LAUNDON, G.F. 1967. Terminology in the Rust
Fungi. Trans. Br. MycoL Soa 60: 189-194.
PARANDEKAR, SA. and S.L. AJREKAR. 1932. Observ-
ations on Pucdnia thwaitesii Berk. Jour, Indian Bot,
Soc. 11: 127-131.
PAYAK, M.M. 1952. Diploidisation in Uromyces hob-
soni Vize and Pucdnia thwaitesii Berk. Indian Phyto-
pathol 5: 104-109.
PETERSEN, R.H. 1974. The Rust Fungus Life Cycle.
The Botanical Review 40: 453-513.
SASS, J.E. 1958. Botanical microtechnique. Ames,
Iowa: Iowa State University Press.
(Received 15 September, 1988)

PERTANIKA VOL. 12 NO. 1, 1989

Pertanika 12(1), 11-14 (1989)

A Rot of Detached Durian Fruits Caused by Sclerotium rolfsii

T.K LIM and KAMARUZAMAN SIJAM
Department of Plant Protection,
Faculty of Agriculture,
Universiti Pertanian Malaysia,
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

Keywords: Durian; Durio zibethinus; Sclerotium rolfsii; fruit rot.

ABSTRAK
Satu penyakit reput buah durian yang baru bagi buah-buahan yanggugur telah diperhatikan di ladang buah-
buahanUniversiti Pertanian Malaysia, Serdang, Selangor, semasa musim durian masak pada bulan Oktober,
1988. Penyebab reput ini membentuk lapisan bebenang miselium kasar yang padat, berwama putih, ber-
bentuk kipas yang tumbuh di atas tompok nekrotik perang berair pada buah yang bersentuhan dengan sisa
rumpai yang mereput di atas tanah. Penyebab penyakit telah dikenalpasti sebagai kulat Sclerotium rolfsii.
Kehadiran pertumbuhan rumpai yang tebal di bawah pokok durian dan sisa daun rumpai yang mereput,
serta keadaan lembab dan panas adalah faktor-faktor perangsang yang penting, yang menggalakkan per-
kembangan reputan tersebut.

ABSTRACT
A new fruit rot of fallen durian fruits was observed in the Uiversiti Pertanian Malaysia fruit orchard in
Serdang, Selangor during the fruit ripening month of October, 1988. The rot was characterized by a dense,
white, fan-shaped mat of coarse my celial strands of the causal fungus growing on a water-soaked, brown necrotic
patch on the fruit in contact with decaying weed vegetation on the ground. The causal organism was identified
as Sclerotium rolfsii. Presence of a thick weed undergrowth and its decaying leaf debris, and warm, moist
conditions were shown to be important predisposing factors conducive to the development of the rot.

INTRODUCTION of durian was encountered on fallen durian

Up to 1980, forty-five different microorganisms
have been listed as associated with multitudi-
nous disorders and diseases of durian, Durio
zibethinus Murr. throughout Peninsular Malaysia
(Johnston 1960, Singh 1980), Sarawak
(Turner 1971), and Sabah (Williams and
Liu 1976). Only one microorganism, the
fungus Rhizopus artocarpi Racib., was listed as as-
sociated with a fruit rot of durian
(Johnston 1960). Recently, a serious fruit rot
of durian was reported to be caused by Phytoph-
thora palmivora Butler (Lim and Chan 1986).
This fruit rot was observed on intact unripe and
ripe fruits and was particularly severe and
rampant in several orchards in Selangor, Pahang
and Johore in Peninsular Malaysia.
Lately, during the fruit ripening month of
October, 1988, a new and interesting fruit rot
fruits in one of the University's durian orchards.
Fruits which had fallen into the thick, profuse
undergrowth of Asystasia intrusa Bl. ('Rumput
Bunga Putih') a weed of the family Acantha-
ceae, around the base of durian trees and left
uncollected for a few days were found to be
infected by a soil-borne fungus. This paper
reports on studies made on the isolation,
identification, and pathogenicity of the fungus.
Predisposing factors influencing the develop-
ment of the fruit rot were also studied.
Symptoms of the fruit rot and methods to avoid
the fruit rot are also elaborated.
MATERIALS AND METHODS
Isolation
Diseased durian fruit tissue ( 5 x 3 x 1 mm) were
taken from the advancing edge with a sterilised
 T.K. LIM AND K. SIJAM
scalpel after removal of the fruit epidermis and
plated on potato dextrose agar (PDA). Mycelial
strands of the fungus growing on the fruit
surface were taken and plated directly on PDA.
The PDA plates were incubated in the dark at
28° C. Hyphal tip transfers were made to obtain
pure cultures from twcwiay old PDA cultures
and pure cultures were maintained on PDA
slants for use in the studies.
Cultural Studies
The following temperature regimes were used
to determine the optimum and cardinal growth
temperatures of the fungus viz. 8, 15, 20, 24,
28, 32, 35, and 40°C. The study was conducted
using a completely randomized design and four
replicate plates were used for each temperature
level. Six mm diameter PDA discs taken from
three-day old cultures were separately placed
centrally on PDA plates and incubated at the
above temperatures. Colony diameters were
measured daily for four days and the data
analysed using analysis of variance. Slides of
the vegetative hyphae of the fungus were pre-
pared for measurement. Sclerotia develop-
ment and measurements were recorded after a
week's incubation.
Pathogenicity Studies
Durian. Mature blemish-free durian fruits
(cv. D 8) freshly fallen from the tree the night
before were used. The surface of the fruit was
surfaced-sterilised with 95% alcohol and inocu-
lated with a 8 mm PDA disc of the test fungus
taken from a four-day old culture. Fruits in the
check treatment were similarly inoculated but
with blank PDA disc without the test fungus.
Inoculated fruits were placed in a) humid
chamber with relative humidity (R.H.) of >95%,
kept in a room of 28°C and b) room with a
temperature of 28°C and R.H. of 75%. Eight
fruits were used for each of the incubation
conditions and the check treatments. Lesion
developments were recorded up to six days after
inoculation. Reisolation of the fungus was done
on PDA.
Three-week-old durian seedlings were
transplanted into sterilised and non-sterilised
soil which were artificially inoculated with
mycelial fragments and sclerotia of the fungus
12 PERTANIKA VOL. 12 NO. 1, 1989
by incorporation into the medium. Another
batch of seedlings transplanted into non-
inoculated, sterilised soil served as check treat-
ment. The seedlings were kept in a moist
chamber at 28°C for three weeks and observed
for development of disease.
Asystasia Intrusa Weeds. Asystasia weeds
growing beneath the durian tree were uprooted
and transplanted into pots filled with sterilised
and non-sterilised soil taken from the durian
orchard. Prior to transplanting, the soils were
inoculated with the test fungus as in the case
of the durian fruit above. Some pots were not
inoculated and served as the check treatments.
All the pots were kept in a humid chamber
at 28°C. The trial was carried out in com-
pletely randomised design and replicated four
times. Disease development was monitored for
two weeks.
In addition, detached Asystasia plant parts
were inoculated with the fungus in moist,
covered, plastic containers and observed for
colonisation of plant parts by the fungus.
RESULTS
Isolation and Cultural Studies
Isolation from diseased durian fruit tissues and
fungal mycelial strands yielded the same fungus.
The fungus was identified as Sclerotium rolfsii
Sacc. - sclerotial state of Corticium rolfsii Curzi
as it exhibited all characteristics of this fungus
as described by Aycock (1966) and Mordue
(1974). It grew readily on PDA, 20-25 mm/day
and attained a diameter of 90 mm within four
days. It produced a white, radiating colony of
tough mycelial strands with a white colony
reverse on PDA. The optimum temperature for
growth was determined to be around 32°C; no
significant difference in growth was found
between 32°C and 28°C (Table 1). No growth
was observed at 8°C and 15°C, and at 40°C the
fungus grew sparingly. Abundant sclerotia (1-2
mm in diameter) were produced in more than
a week old PDA cultures. The sclerotia were
white at first, turning light brown and subse-
quently dark brown, and each had a character-
istic thick rind with distinct medulla and cor-
tex. Clamp connections were observed in the
primary hyphae. Diameter of primary hyphae
varied from 4-7 |im, while tertiary branches were
 A ROT OF DETACHED DURIAN FRUITS CAUSED BY SCLEROTIUM ROIfSII

much narrower, 2 |im. The teleomorphic state"

was not observed.
TABLE 1
Effect of temperature on colony growth of
Sderotium rolfsii

Temperature (°C) Mean colony diameter (cm)

8 0 g1
15 0.39 f
20 1.57 d
25 L80 be
28 1.86 a
32 1.89 a Fig. 2. Internal necrosis caused by S. rolfsii extending
35 1.73 c into the seed cavity, rotting the seed and discolour-
40 0.47 e ing the cotyledons (arrowed).
df = 14
MSE = 0.01639 The pathogenicity studies demonstrated
LSD = 0.07 that S. rolfsii was able to infect mature durian
1 fruits without prior wounding, producing similar
Mean values followed by similar letters denote no sig-
nificant difference at P = 0.05 as determined by LSD
symptoms as those observed on fallen fruits in
test the thick Asystasia undergrowth. Fruits kept at
both levels of relative humidity i.e. 75 and >95%
Pathogenicity Studies at 28°C developed the rot. However, those
In the durian field, this peculiar fruit rot could incubated at the higher R.H. had more exten-
easily be distinguished by the white, fan-shaped sive rotting and more luxuriant growth of the
tufts of thick mycelial strands of the fungus fungus on the fruits. The fungus was easily
growing on the water-soaked, brown, necrotic isolated from infected fruits, thus satisfying
lesion on the durian fruit {Fig. 1) in contact Koch's postulates. The fungus did not infect
with decaying Asystasia weed debris on the the durian seedlings in both sterilised and non-
ground. The weed debris was caused by the sterilised soil.
durian fruit falling amongst the thick Asystasia S. rolfsii grew veiy well on the detached
undergrowth. Internally, necrosis extended into Asystasia plant parts but did not infect intact,
the seed cavity, rotting the seed and discolou- healthy Asystasia plants in both sterilised and
ring its cotyledons [Fig. 2). nonsterilised soils.

Fig. L Tufts of thick, white, fan-shaped mycelial strands
o/Sclerotium rolfsii covering the brown, water-
soaked, necrotic lesion on an infected durian fruit.
DISCUSSION
S. rolfsii is an ubiquitous fungus, found grow-
ing in a diverse array of soil types in the
tropics and subtropics (Abeygunawardena and
Wood 1957; Aycock 1966). In the soil it
survives by means of its sclerotia and thick
mycelial strands, growing saprotrophically on
decaying plant debris and organic matter
(Mordue 1974). Aycock (1966) reported that
the fungus causes root and stem rot and other
diseases in about 500 plant species from 100
families, most of which are dicotyledons com-
prising mainly composites and legumes. It is a
facultative parasite, but has not been recorded
to attack durian or its plant parts before. This

PERTANIKA VOL. 12 NO. 1, 1989

 T.K. LIM AND K. SIJAM
represents the first record of its facultative
parasitism on durian.
The pathogenicity results indicate that
durian is not a natural host of S. rolfsii How-
ever, once its fruits became detached and fell
on the ground where the fungus was present,
infection of the fruit occurred especially under
warm and moist conditions as was demonstrated
in the pathogenicity tests on the fruits. Such a
conducive environment in the orchard could
be provided by the thick undergrowth of
Asystasia weeds around the tree. Empirical
evidence suggests that the falling durian fruit
damaged some of the thick weed undergrowth
and the damaged weed debris provided a ready
food source for the luxuriant saprotrophic
growth of the fungus. The fallen fruit was left
in situ amongst the thick undergrowth, uncol-
lected for a few days as it was hidden from the
picker's view. Such a condition predisposed the
fruit to infection by the fungus which was already
established on the weed debris. This was also
shown by infection studies on the weed: the
fungus grew well on the detached Asystasia
plant parts under warm, moist conditions
but did not infect the healthy Asystasia under
similar conditions.
To avoid Sclerotium rot of durian fruits,
thorough and frequent (once or twice daily)
picking of fallen fruits should be carried out.
The base of the durian tree should be kept free
from thick weed undergrowth which creates a
warm and moist condition for infection of fallen
fruits by soil fungi. Also, by having a thick
undergrowth of weeds beneath the durian tree,
many fallen fruits will remain unpicked as they
will be obscured by the dense undergrowth.
Besides, weeds compete with the tree for
nutrients. Another way is to erect a strong canvas,
nylon or wire netting beneath the durian tree
to break the fall of the ripe fruits and to reduce
fruit damage and mitigate the risk of infection
PERTANIKA VOL. 12 NO. 1, 1989
by soil fungi. This can be a practical proposi-
tion especially with high priced cultivars like D
2, D 24, D 98, D 99 and D 114 where each fruit
can fetch a price of more than $12 ringgit in
the retail market.
ACKNOWLEDGMENT
The authors express their thanks to the Deputy
Director of the University Farm for providing
the fruits used in this study and the MPKSN
Council for a grant on fruit diseases.
REFERENCES
ABEYGUNAWARDENA, D.V.W. and R.K.S. WOOD. 1957.
Factors Affecting the Germination of Sclerotial
and Mycelial Growth of Sclerotium rolfsii. Trans.
Brit. Mycol. Soc. 40(2): 821-251.
AYCOCK, R. 1966. Stem Rot, and Other Diseases
Caused by Sclerotium rolfsii. North Carolina Agric.
Expt. Stat. 174, 202 pp.
JOHNSTON, A. 1960. A Supplement to a Host List
of Plant Diseases in Malaya. Mycological Paper No,
77. Commonwealth Mycological Institute, Kew,
England.
LIM, T.K. and L.G. CHAN. 1986. Fruit Rot of
Durian Caused by Phytophthora palmivora. Pertanika
9(3): 269-276.
MORDUE, J.E.M. 1974. Corticium rolfsii. CMI
Description of Pathogenic Fungi and Bacteria
No. 410.
SINGH, KG. 1980. A Check List of Host and
Diseases in Malaysia. Ministry of Agric. Malaysia
Bull. No. 154.
TURNER, GJ. 1971. Fungi and Plant Disease in
Sarawak. Phytopaihological Paper No. 13, Commmon-
wealth Mycological Institute, Kew, England.
W I L L I A M S X H . and P.S.W. Liu. 1976. A host list
of Plant Disease in Sabah. Phytopathological Paper
No. 19, Commonwealth Mycological Institute, Kew,
England.
(Received 29 November, 1988)
Pertanika 12(1), 15-25 (1989)

Nutrient Composition of
Selected Cooked and Processed Snack Foods
E.S. TEE, S. SITI MIZURA, A. ANUAR,1 R. KULADEVAN,
S.I. YOUNG, S.C. KHOR, and S.K. CHIN.
Division of Human Nutrition
Institute for Medical Research, 50588 Kuala Lumpur, Malaysia

ABSTRAK
Kandungan zat di dalam 27 makanan ringan yang dimasak dan 19yang diproses telah dikaji. Kebanyakan
makanan yang dimasak itu berasaskan bijirin, dibuat daripada tepung gandum, beras dan tepung beras,
dan hampir kesemuanya adalah kuih atau masakan tradisional Malaysia. Makanan ringan yang diproses
terdiri dart hasil-hasil coklat, bijirin, ikan dan udang. Kandungan 19 zat dalam setiap WOg bahagian yang
boleh dimakan telah dibentangkan. Kandungan zat yang terpilih dalam setiap hidangan atau bungkusan
makanan juga dibentangkan. Laporan ini bermaksud untuk membantu menambahkan pengetahuan mengenai
data komposisi makanan ringan tempatan. Bilangan makanan yang telah dikaji hanyalah sebahagian kecil
daripada jumlah yang sedia ada. Memandangkan data seperti ini amat berkurangan, lebih banyak kajian
perlu dijalankan dalam bidang ini untuk memenuhi permintaan data yang kian meningkat.

ABSTRACT
Nutrient composition of 27 cooked snack foods and 19 processed snacks was determined. The cooked foods were
mostly cereal based, made from wheat flour, rice or rice flour, and almost all of them were traditional Malaysian
kuih or dishes. The processed snacks studied were chocolate, cereal, tuber, fish and prawn products. The levels
of 19 nutrients were tabulated, expressed as per 100 g edible portion. Selected nutrients in each serving or
packet of the foods were also presented. The paper is intended as a contribution to the knowledge on nutrient
composition of local snack foods, for which information is still greatly lacking. The number of foods studied
is only a fraction of the total number available. More work in this area will have to be carried out, to meet
the increasing demand for such data.

INTRODUCTION been more studies on these foods. In fact,

The role of snack foods in the nutrition of
children in Malaysia has recently been given
considerable attention. Tee (1979) had empha-
sized the importance of selling nutritious snack
foods in school canteens. Zanariah (1986) dealt
with some aspects of the controversy surround-
ing the nutritive value of snack foods. Aminah
etal (1987) reported the consumption of snack
foods by rural primary school children in Kedah.
Various newspapers and magazines have also
highlighted the issue from time to time.
Data on the nutritive value of snack foods
is lacking. The food composition table for use
in Malaysia (Tee, 1985) also does not provide
figures for these foods. Of late, there has
'Food Quality Control Unit, Ministry of Health Malaysia, Kuala Lumpur.
there has been greater emphasis on study-
ing the nutrient composition of cooked foods
in general, to meet the increasing demand
for such data (Tee et al. 1986). The works
of Tee et al (1979) and Mohamad Nordin
(1983) provide some data on Malaysian
cooked foods, whereas more recently, Zanariah
(1986) and Aminah et al (1987) reported
the level of some nutrients in selected
snack foods.
This report is aimed at contributing to
the knowledge on food composition data of
local foods. The composition of 19 nutrients of
over 40 types of commonly consumed snack
foods is reported.
 E.S. TEE, S. SITI MIZURA, A. ANUAR, R. KU1ADEVAN, S.I. YOUNG, S.G KHOR AND S.K. CHIN.
MATERIALS AND METHODS
A total of 27 samples of cooked snack foods and
19 processed snacks were studied. Most of the
foods were obtained from several school
canteens in Selangor Darul Ehsan. One sample
of each of the cooked foods was taken for
analysis, except for cooked meals with different
ingredients. For the latter, at least two samples
were taken from different canteens and analysed
separately. All cooked foods were collected and
brought to the laboratory the same morning.
For the processed snacks, where appropriate,
several brands of the foods were collected. The
weight of the edible portion of each serving, or
piece, or packet, as the case may be, was re-
corded. The ingredients of the cooked foods
were noted, or enquired from the vendor. For
the processed snacks, the ingredients were taken
from the packaging. A portion of each food was
blended for immediate determination of ascor-
bic acid. The remainder was blended and
aliquots weighed out for the various analyses,
carried out in duplicate.
All analyses were performed according to
the methods given in the laboratory manual
compiled by the Institute for Medical Research
(Tee et al 1987). Moisture was determined by
the air-oven method; protein by the semi-micro
Kjeldahl method; fat was extracted using the
Soxhlet apparatus; crude fibre by the acid-alkali
digestion method, and ash content was deter-
mined after incinerating the food in a muffle.
Carbohydrate content was then determined by
subtracting from 100 the content of moisture,
protein, fat, crude fibre and ash. Energy con-
tent of the food was calculated by multiplying
the protein, carbohydrate and fat values with
the Atwater factors of 4,4 and 9 respectively.
Five minerals were determined on the ash
solution prepared from the food. Calcium was
determined by titration against potassium
permanganate. Phosphorus was determined
colorimetrically after reaction with vanadate-
molybdate reagent, and iron content was
similarly estimated colorimetrically through its
reaction with o-phenanthroline. Sodium and
potassium were determined using atomic
absorption spectrometry.
Vitamin A and carotene were first ex-
tracted from the unsaponifiable fraction of the
16 PERTANIKA VOL. 12 NO. 1, 1989
saponification mixture and then chromato-
graphed on a column of alumina. The pro-
vitamin and vitamin A thus separated were read
separately in a spectrophotometer. Total vitamin
A activity in the food was calculated from the
sum of retinol concentration and 1/6 the con-
centration of carotene (WHO 1967), and ex-
pressed as jug retinol equivalent (RE). Thiamine
was determined using the thiochrome pro-
cedure, whilst riboflavin was estimated in a
fluorometer. Niacin concentration was colori-
metrically determined after reaction with
cyanogen bromide and sulphanilic acid. The
indophenol dye titration method was used for
the estimation of ascorbic acid content of
the foods.
RESULTS AND DISCUSSION
Types of Snack Foods and Ingredients
The 27 cooked snack foods analysed are listed
in Table 1. Ingredients of the foods studied are
also given in the table. Where appropriate, the
method of preparation of the food is also
indicated. The word 'snack' is used in a broad
sense, and the list includes 5 types of foods
which may be considered as meals in some
context. These include various preparations of
mee, mee-hoon, kuih-tiau and nasi lemak. Different
samples of the same cooked meal were
separately listed since the ingredients used were
different. Based on the main ingredient, the
foods studied were divided into three groups.
The first group consisted of eleven wheat flour
based snacks, and the second group was made
up of another eleven foods using rice or rice
flour as the main ingredient. Three items based
on banana, and two on legumes were placed in
the third group. Even in this miscellaneous
group, wheat or rice flour was an important
ingredient in four out of the five foods studied.
Coconut or santan was another commonly used
ingredient, featured in at least 12 of all the
foods studied. Almost all the cooked snack foods
examined in this study were traditional
Malaysian kuih or foods based on local recipes.
Hence, their local names are used in this paper.
Table 2 lists the 19 processed snacks
analysed. The names of the snacks and the
ingredients listed are as given on the wrapper.
They have been loosely grouped as chocolate,
 NUTRIENT COMPOSITION OF SELECTED COOKED AND PROCESSED SNACK FOODS

TABLE 1
Ingredients of cooked snack foods

Food Ingredients/Method of preparation

Wheat flour based:

Cake, plain, cup Flour, margerine, egg, sugar, baking powder; baked
Cucur badak Flour, coconut, dried prawn, salt, chilli; deep fried in oil
Currypuff Flour, potatoes, spices, spring onion, curry leaves; deep fried in oil
Kuih apam Flour (and rice flour), sugar, yeast, salt, coconut; steamed
Kuih kasui Flour, brown sugar, salt, coconut, lime-water; steamed
Kuih keria Flour, sweet potato, sugar, salt; deep fried in oil
Kuih udang Flour, dried prawn, salt, kuchai, bean sprout; deep fried in oil
Sandwich, sardine Bread (plain), sardine (canned), ikan bilis
Fried mee (sample 1) Mee, bean sprout, egg, mustard leaf; fried
Fried mee (sample 2) Mee, ikan bilis, kangkong, onion, chilli; fried
Mee soup Mee, bean sprout, fish cake, tau-pok, parsley, sambal, onion; boiled
Rice/rice flour based:
Buah Melaka Glutinous rice flour, gula melaka, coconut, salt, colouring; boiled
Kuih sri muka Glutinous rice, santan, sugar, flour, salt, colouring; steamed
Pulul panggang/pulut udang Glutinous rice (boiled), coconut, dried prawn, onion, garlic, dried chilli,
serai; grilled
Kuih lapis Rice flour, sugar, santan, salt, colouring; steamed
Fried kuih-liau Kuih-tiau, lau-kua, bean sprout, eggs, fish cake, chilli, ketchup; fried
Fried mee-hoon (sample 1) Mee-hoon, bean sprout, kuchai, strips of fried egg, chilli; fried
Fried mee-hoon (sample 2) Mee-hoon, ikan bilis, kangkong, sambal, onion; fried
Fried mee-hoon (sample 3) Mee-hoon, fishball, mustard leaf; fried
Nasi lemak (sample 1) Rice with santan (steamed), ikan bilis, egg, cucumber, chili
Nasi lemak (sample 2) Rice with santan (steamed), ikan bilis, sambal
Nasi lemak (sample 3) Rice with santan (steamed), ikan bilis, sambal, cucumber, onion
Miscellaneous:
Cokodok pisang Banana, flour, sugar, salt; deep fried in oil
Ijepal pisang Banana, flour, santan, sugar, salt; steamed in banana leaves
Pisang goreng Banana, wheat or rice flour; deep fried in oil
Kuih kasluri Green bean, rice flour, coconut, egg, sugar, salt; deep fried in oil
Bubur kacang merah Red beans, sugar; boiled

cereal and tuber, and fish and prawn pro-
ducts, based on their main ingredients. The
most popularly consumed are probably the
cereal and tuber based products, or the ex-
truded snack foods.
Nutrient Composition of Crooked Snack Foods
Proximate composition of the cooked snack
foods, expressed as per 100 g of edible portion,
is tabulated in Table 3. Most of the foods were
found to have a rather narrow range of energy
level of 130 to 255 Kcal per 100 g edible por-
tion, except for cake and the two soup dishes
(mee soup and bubur kacang merah) which had
a moisture content of over 80%. Protein con-
centration also did not vary widely, with most
of the values ranging from about 2 to 5 g. The
three exceptions were cake, sardine sandwich
and kuih kasturi, which had protein level of
about 7 g. The cooked foods studied did not
have a high fat content, most of them below
9 g, with a median value of 3.3 g. The except-
ions were cake and currypuff. Being mostly
cereal based foods, they were found to have a
fairly high carbohydrate content. Leaving out
the two soup dishes, carbohydrate level was
found to range from about 23 to 58 g, with a
median value of 35 g per 100 g edible portion.
Table 4 tabulates the mineral content of
the cooked snack foods, expressed as per 100 g
edible portion. Except for 6 items which had a
calcium level of below 10 mg, the majority of

PERTANIKA VOL. 12 NO. 1, 1989 17

 E.S. TEE, S. SITI MIZURA, A. ANUAR, R. KULADEVAN, S.I. YOUNG, S.C. KHOR AND S.K. CHIN.

the other foods had around 10 to 30 mg per
100 g edible portion. Sardine sandwich had an
exceptionally high calcium level of 94 mg. Kuih
kasturiwa.s found to have an exceptionally high
concentration of iron, whilst the others had
levels ranging from 0.2 to 2 mg.
The vitamin content of the cooked foods
is tabulated in Table 5. Some of the foods were
found to have moderate quantities of vitamin
A activity, higher than for cereals in general.
These are foods that contained animal products
as ingredients, and a higher proportion of the
vitamin A activity was from retinol. Thiamine
and niacin levels in the foods were found to 1
rather low, compared with levels found in ric
rice flour and wheat flour. Only one of tl
snacks, kuih kasturi, had a vitamin B, level i
more than 0.1 mg per 100 g sample. On tr
other hand, riboflavin was not exceptional
low, compared with the cereals mentions
above.
In order to provide a quick reference 1
the composition of selected nutrients in eac
serving of the cooked snack foods, these ha\
been calculated and tabulated in Table <
Portion sizes of the foods showed a gener;

TABLE 2
Ingredients of processed snack foods

Food1 Ingredients1

Chocolate products:
Bola coklat bertriskut
Chocolate wafer (brand 1)
Chocolate wafer (brand 2)
Milk chocolate beans
Milk chocolate peanuts
Cereal and tuber products:
Corn stick, chocolate flavour
Noodle snack, chicken flavour
Pop Corn, durian flavour
Snacks, cheese flavour
Snacks, chicken flavour (brand 1)
Snacks, chicken flavour (brand 2)
Snacks, chicken flavour (brand 3)
Potato chips (brand 1)
Potato chips (brand 2)
Fish and prawn products:
Fish 'satay'
Prawn crackers
Cuttlefish crackers
Prepared cuttlefish (brand 1)
Prepared cuttlefish (brand 2)
Cocoa powder, edible vegetable oil, sugar, flour
Cocoa, cocoa mass milk, butter, sugar, starch, wheat flour
Cocoa powder, sugar, flour, milk powder, vegetable oil, vanilir
permitted flavouring and colouring
Cocoa powder, sugar, glucose, permitted colouring
Cocoa, peanuts, vegetable fats, milk powder, 'lincintin' soyi
sugar
Corn, rice, edible oil, cocoa, cream, salt
Wheat flour, palm oil, salt, potato starch, sodium carbodymethy
cellulose, sodium and potassium carbonate, permitted coloui
monosodium glutamate, soya sauce, spices and chicken essenc
Corn, sugar, salt, margarine, milk powder, flavouring an<
permitted colouring
Maize, cheese, edible vegetable oil, spices, permitted colourinj
Corn, edible vegetable oil, artificial chicken flavouring, permittee
colouring
Rice, corn, edible vegetable oil, artificial chicken flavouring
permitted colouring
Rice, maize, edible vegetable oil, artificial chicken flavouring
permitted colouring
Potato, vegetable oil, salt and approved flavourings
Potato flour, wheat flour, vegetable oil, permitted flavouring
and colouring
Fish, starch, salt, sugar, chillies
Fresh prawns, wheat flour, vegetable oil, modified food starch
monosodium glutamate, salt
Cuttlefish, wheat flour, vegetable oil, starch, salt, monosodium
glutamate
Cuttlefish, sugar, chillie, pepper, salt, monosodium glutamate
Not stated

'Name of food and ingredients are as given on the wrapper

18 PERTANIKA VOL. 12 NO. 1, 1989

 NUTRIENT COMPOSITION OF SELECTED COOKED AND PROCESSED SNACK FOODS

trend, in that the kuih and similar snacks
weighed around 15 to 70 g per serving, whereas
for the noodle dishes and nasi lemak, serving
weights ranged from 100 to 200 g. Neverthe-
less, there were considerable differences in
portion sizes, and hence, there was more vari-
ation in the level of most of the nutrients listed.
Nutrient Composition of Processed Snack Foods
Table 7 tabulates the proximate composition of
the processed snacks analysed, expressed as per
100 g edible portion of the foods. It can be seen
that there is considerable variation in the protein
and fat contents of the foods. Carbohydrate
level showed less variation, although an ex-
tremely low level was obtained for one of the
prepared cuttlefish products. Energy level of
the foods varied within a rather narrow range
of between 300 to 500 Kcal per lOOg.
Levels of the 5 minerals determined are
tabulated in Table 8. A considerable amount of
variation is observed, and there appears to be
no general trend, even for items within each
sub-group. This is also true for the vitamins
(Table 9). Contributions to total vitamin A
activity were almost solely from retinol for
the chocolate, fish and prawn products. In the
case of cereal products, carotene also con-

TABLE3
Proximate composition of cooked snack foods (per lOOg edible portion)
Food1 Energy Moisture Protein Fat Carbo- Crude Ash
hydrate fibre
Kcal % g g g g
Wheat flour based:
Cake, plain, cup 435 14.8 7.1 19.8 57.1 0 1.2
Cucur badak 238 46.9 5.5 6.6 39.2 0.7 1.1
CurrypufF 330 33.6 4.7 14.7 44.8 0.3 1.9
Kuih apam 193 52.0 2.4 0.4 45.0 0 0.2
Kuih kasui 131 67.2 2.5 0.5 29.1 0.4 0.3
Kuih keria 236 44.2 2.6 3.9 47.7 0.8 0.8
Kuih udang 248 47.5 5.5 8.7 36.9 0.1 1.3
Sandwich, sardine 252 41.7 7.4 5.8 42.6 0.6 1.9
Fried mee (sample 1) 161 60.2 5.5 3.3 27.2 0.1 1.7
Fried mee (sample 2) 148 62.4 4.8 1.4 29.0 0.3 2.1
Mee soup 68 84.9 3.6 2.6 7.5 0.2 1.2
Rice/rice flour based:
Buah Melaka 201 50.3 3.2 1.2 44.4 0.5 0.8
Kuih sri muka 177 58.3 3.5 2.6 35.0 0 0.6
Pulut panggang/pulut udang 219 48.6 4.0 4.6 40.5 1.5 0.8
Kuih lapis 146 63.7 3.0 0.6 32.2 0 0.5
Fried kuih-tiau 170 63.9 5.7 6.4 22.5 0.1 1.4
Fried mee-hoon (sample 1) 138 69.7 2.9 4.0 22.5 0.3 0.6
Fried mee-hoon (sample 2) 166 59.0 3.6 1.2 35.1 0.2 0.9
Fried mee-hoon (sample 3) 158 64.7 3.9 3.9 26.7 0.2 0.6
Nasi lemak (sample 1) 165 62.8 5.6 4.9 24.6 0.5 1.6
Nasi lemak (sample 2) 162 61.7 4.6 2.5 30.2 0.2 0.8
Nasi lemak (sample 3) 149 64.8 4.1 2.6 27.2 0.3 1.0
Miscellaneous:
Cokodok pisang 254 43.2 3.9 6.3 45.4 0.5 0.7
Lepat pisang 158 60.6 3.2 0.7 34.8 0 0.7
Pisang goreng 184 55.3 1.8 2.1 39.5 0.3 1.0
Kuih kasturi 246 45.1 6.9 7.3 38.2 1.5 1.0
Bubur kacang merah 44 88.2 1.6 0.1 9.2 0.7 0.2
'No refuse in all samples

PERTANIKAVOL. 12 NO. 1, 1989

 E.S. TEE, S. SITI MIZURA, A. ANUAR, R. KULADEVAN, S.I. YOUNG, S.C. KHOR AND S.K. CHIN.

tributed to total vitamin A. As was d o n e for the
cooked foods, the composition of selected
nutrients in each packet or serving of the
processed snacks have been calculated and
tabulated in Table 10.
CONCLUSION
No a t t e m p t has been m a d e to c o m p a r e the
nutritive value of the two groups of snack foods
studied, cooked and processed. Firstly, the
n u m b e r of items studied was too small for a
valid comparison. In any case, the nutrients
derived from a particular food d e p e n d on the
quantity c o n s u m e d . F u r t h e r m o r e , the cost of
deriving a quantity of a certain nutrient will
have to be taken into consideration.
It is also not the intention of this r e p o r t
to deal with the controversy of t h e c o n s u m p -
tion of snack foods, especially a m o n g children.
It is h o p e d that the results will provide some
u n d e r s t a n d i n g on the n u t r i e n t composition of
these foods. In no way does the r e p o r t claim to
provide representative data for the n u t r i e n t
composition of the foods studied, since no
p r o p e r sampling p r o c e d u r e was carried out. It
does provide an idea of the o r d e r of m a g n i t u d e
of the nutrients tabulated for each type of food.
It must, however, be b o r n e in m i n d that con-
siderable variation in nutrient c o n t e n t could
exist a m o n g the cooked foods, d e p e n d i n g
mainly on the composition of t h e ingredients.
This is in fact a major problem with t h e analysis

TABLE 4
Mineral content of cooked snack foods (mg per lOOg edible portion)

Food Calcium Phosphorus Iron Sodium Potassium

Wheat flour based:

Cake, plain, cup 15 160 0.8 356 38
Cucur badak 27 59 0.8 145 183
CurrypufF 26 54 1.3 127 109
Kuih apam 5 27 0.3 844 0
Kuih kasui 55 17 0.4 42 17
Kuih keria 29 59 0.3 81 164
Kuih udang 21 54 0.6 376 41
Sandwich, sardine 94 80 1.6 457 93
Fried mee (sample 1) 13 33 0.8 550 44
Fried mee (sample 2) 27 44 1.1 595 55
Mee soup 15 27 0.5 302 29
Rice/rice flour based:
Buah Melaka 31 29 2.3 153 59
Kuih sri muka 8 29 0.5 146 37
Pulut panggang/pulut udang 11 40 0.4 171 94
Kuih lapis 7 34 0.4 110 34
Fried kuih-tiau 20 53 2.0 366 33
Fried mee-hoon (sample 1) 16 34 0.8 156 38
Fried mee-hoon (sample 2) 34 39 2.0 192 31
Fried mee-hoon (sample 3) 19 16 0.7 92 24
Nasi lemak (sample 1) 41 73 1.2 360 60
Nasi lemak (sample 2) 13 40 0.6 180 48
Nasi lemak (sample 3) 23 50 0.5 251 59
Miscellaneous:
Cokodok pisang 9 46 0.6 75 162
l^pat pisang 8 21 0.4 57 141
Pisang goreng 7 39 0.6 108 216
Kuih kasturi 28 110 3.4 95 185
Bubur kacang merah 8 14 0.2 3 40

20 PERTANIKA VOL. 12 NO. 1, 1989

 NUTRIENT COMPOSITION OF SELECTED COOKED AND PROCESSED SNACK FOODS

of cooked foods in general. Nevertheless, for
many cooked foods, the portion size and ingre-
dients of a particular food from different
vendors are rather similar. To enhance the use-
fulness of data on cooked foods, particular at-
tention will also have to be paid to the listing
of ingredients in the food or dish and method
of preparation. There has been an increasing
demand for data on the nutrient composition
of cooked foods, and more intensive work in
this area will have to be carried out.
The snack foods analysed in the study are
only a fraction of the total number of these
foods available. There are many more interest-
ing and delicious Malaysian cooked snacks.
Some of them are described in the various recipe
and cook books. A comprehensive compilation
of all these foods, with their recipes and method
of preparation should be useful. For the pro-
cessed snacks, the industry has been growing
rapidly, and knowledge on the nutrient com-
position of snack foods is still limited.
ACKNOWLEDGEMENTS
The authors thank the Director of the Institute
for Medical Research, for permission to publish

TABLE 5
Vitamin content of cooked snack foods (per 100 g edible portion)

Food Retinol Carotene Vitamin A Thiamine Riboflavin Niacin Ascorbic

Activity Acid
HgRE1 mg mg mg mg
w
Wheat flour based:
Cake, plain, cup 65 0 65 0.06 0.05 0.6 1.0
Cucur badak 71 64 82 0.09 0.04 0.7 0.9
Currypuff 40 54 49 0.07 0.04 0.8 1.2
Kuih apam 4 0 4 0.03 0.2 0
Kuih kasui 8 0 8 0.03 0.3 3.8
Kuih keria 19 57 29 0.08 0.04 0.8 0
Kuih udang 34 35 40 0.07 0.04 1.3 0.6
Sandwich, sardine 46 169 74 0.09 0.07 1.1 1.1
Fried mee (sample 1) 11 23 14 0.01 0.04 0.9 1.1
Fried mee (sample 2) 43 97 59 0.01 0.04 0.4 5.7
Mee soup 14 0 14 o 0.02 0.4 0

Rice/rice flour based:

Buah Melaka 3 0 3 0 0.05 0.4 1.2
Kuih sri muka 11 0 11 0.01 0.05 0.5 2.0
Pulut panggang/pulut udang 14 72 26 0.02 0.08 1.0 0
Kuih lapis 5 0 5 0 0.05 0.4 1.5
Fried kuih-tiau 43 18 46 0.05 0.08 0.2 1.8
Fried mee-hoon (sample 1) 32 87 47 0.02 0.08 0.3 0.1
Fried mee-hoon (sample 2) 39 266 83 0.01 0.04 0.4 1.7
Fried mee-hoon (sample 3) 26 268 70 * 0.01 0.10 1.9 1.7
Nasi lemak (sample 1) 27 39 34 0.06 0.06 1.2 0.6
Nasi lemak (sample 2) 19 20 22 0.03 0.04 0.8 1.1
Nasi lemak (sample 3) 12 14 14 0.01 0.04 0.5 0.6
Miscellaneous:
Cokodok pisang 31 35 37 0.06 0.06 0.6 0.9
l^epat pisang 8 201 42 0.06 0.07 3.2 0
Pisang goreng 15 39 22 0.05 0.05 0.8 6.5
Kuih kasturi 20 27 25 0.13 0.07 1.8 0
Bubur kacang merah 10 0 10 0 0.02 0.4 0.7

'RE = retinol equivalent

PERTANIKA VOL. 12 NO. 1, 1989 21

 E.S. TEE, S. SITI MIZURA, A. ANUAR, R. KULADEVAN, S.I. YOUNG, S.C. KHOR AND S.K. CHIN.

the results of the study. The assistance of the Darul Ehsan in obtaining samples for the
various local health authorities in Selangor analysis is greatly appreciated.

TABLE 6
Selected nutrients in each serving of cooked snack food

Food Weight Energy Protein Calcium Iron Vitamin A Thiamine Ribo-

activity flavin
g Kcal g mg mg Ug RE1 mg mg

Wheat flour based:

Cake, plain, cup 23 100 1.6 3 0.2 15 0.01 0.01
Cucur badak 29 69 1.6 8 0.2 24 0.03 0.01
CurrypufF 22 73 1.0 6 0.3 11 0.02 0.01
Kuih apam 27 52 0.6 1 0.1 1 0 0.01
Kuih kasui 62 81 1.6 34 0.2 5 0 0.02
Kuih keria 38 90 1.0 11 0.1 11 0.03 0.02
Kuih udang 30 74 1.7 6 0.2 12 0.02 0.01
Sandwich, sardine 28 71 2.1 - 26 0.4 21 0.03 0.02
Fried mee (sample 1) 111 179 6.1 14 0.9 16 0.01 0.04
Fried mee (sample 2) 167 247 8.0 45 1.8 9S 0.02 0.07
Mee soup 563 382 20.3 82 2.8 79 0 0.11
Rice/rice flour based:
Buah Melaka 15 30 0.5 5 0.3 0 0 0.01
Kuih sri muka 53 94 1.9 4 0.3 6 0.01 0.03
Pulut panggang/pulut udang 74 162 3.0 8 0.3 19 0.01 0.06
Kuih lapis 55 80 1.7 4 0.2 3 0 0.03
Fried kuih-tiau 200 341 . 11.4 39 4.0 92 0.10 0.16
Fried mee-hoon (sample 1) 200 275 5.8 31 1.6 94 0.04 0.16
Fried mee-hoon (sample 2) 101 168 3.6 34 2.0 84 0.01 0.04
Fried mee-hoon (sample 3) 186 294 7.3 35 1.3 130 0.02 0.19
Nasi lemak (sample 1) 150 247 8.4 62 1.8 51 0.09 0.09
Nasi lemak (sample 2) 116 188 5.3 15 07 26 0.03 0.05
Nasi lemak (sample 3) 172 256 7.1 40 0.9 24 0.02 0.07
Miscellaneous:
Cokodok pisang 71 180 2.8 7 0.4 26 0.04 0.04
Ijepat pisang 90 142 2.9 7 0.4 38 0.05 0.06
Pisang goreng 65 120 1.2 5 0.4 14 0.03 0.03
Kuih kasturi 72 177 5.0 20 2.4 18 0.09 0.05
Bubur kacang merah 230 101 3.7 17 0.5 23 0 0.05
'RE= retinol equivalent

22 PERTANIKA VOL. 12 NO. 1, 1989

 NUTRIENT COMPOSITION OF SELECTED COOKED AND PROCESSED SNACK FOODS

TABLE 7
Proximate composition of processed snack foods (per lOOg edible portion)

Food1 Energy Moisture Protein Fat Carbo- Crude Ash

hydrate fibre
Kcal g 8 g g g
Chocolate products:
Bola coklat berbiskut 517 2.6 10.1 26.6 59.2 0 1.5
Chocalate wafer (brand 1) 515 1.7 4.6 25.1 67.6 0 1.0
Chocolate wafer (brand 2) 565 2.3 5.5 35.7 55.4 0 1.1
Milk chocolate beans 377 2.5 2.2 2.6 86.2 0 6.5
Milk chocolate peanuts 509 6.8 17.2 30.9 40.5 2.6 2.0
Cereal and tuber products:
Corn stick, chocolate flavour 526 3.9 5.2 28.8 61.5 0 0.6
Noodle snack, chicken flavour 456 3.5 11.0 15.6 67.9 0 2.0
Pop corn, durian flavour 380 4.8 6.9 0.8 86.4 0 1.1
Snacks, cheese flavour 499 3.1 9.3 24.0 6L4 0 2.2
Snacks, chicken flavour (brand 1) 504 3.7 6.2 25.6 62.1 0 2.4
Snacks, chicken flavour (brand 2) 458 6.2 8.0 18.4 65.2 0 2.2
Snacks, chicken flavour (brand, 3) 456 6.4 4.8 17.9 68.8 0 2.1
Potato chips (brand 1) 552 5.6 6.4 39.0 43.9 1.7 3.4
Potato chips (brand 2) 407 7.0 6.3 10.5 71.9 0 4.3
Fish and prawn products:
Fish 'satay' 380 15.2 13.9 12.2 53.7 0.8 4.2
Prawn crackers 424 3.3 6.2 9.8 77.7 0 3.0
Cuttlefish crackers 429 4.1 5.2 10.7 77.9 0 2.1
Prepared cuttlefish (brand 1) 353 9.0 37.5 0.8 49.0 0.2 3.5
Prepared cuttlefish (brand 2) 296 20.9 62.1 3.1 4.9 0.2 8.8
'No refuse in all samples

TABLE 8
Mineral content of processed snack foods (mg per 100 g edible portion)

Food Calcium Phosphorus Iron Sodium Potassium

Chocolate products:
Bola coklat berbiskut 166 189 2.2 86 309
Chocolate wafer (brand 1) 89 98 3.8 87 175
Chocolate wafer (brand 2) 157 97 1.5 98 95
Milk chocolate beans 54 38 9.0 36 117
Milk chocolate peanuts 152 245 2.5 83 295
Cereal and tuber products:
Corn stick, chocolate flavour 29 71 1.4 18 83
Noodle snack, chicken flavour 20 135 1.1 613 88
Pop Corn, durian flavour 18 71 1.5 282 85
Snacks, cheese flavour 111 114 3.3 483 55
Snacks, chicken flavour (brand 1) 14 51 2.6 703 46
Snacks, chicken flavour (brand 2) 8 61 0.8 465 53
Snacks, chicken flavour (brand 3) 6 64 0.8 546 15
Potato chips (brand 1) 17 104 1.4 272 771
Potato chips (brand 2) 34 156 1.0 1152 41.
Fish and praxvn products:
Fish 'satay' 81 175 4.1 841 358
Prawn crackers 94 98 1.1 750 116
Cuttlefish crackers 46 81 1.8 527 110
Prepared cuttlefish (brand 1) 51 364 4.6 923 271
Prepared cuttlefish (brand 2) 143 717 6.5 1352 585

PERTANIKA VOL. 12 NO. 1, 1989

 E.S. TEE, S. SITI MIZURA, A. ANUAR, R. KULADEVAN, S.I. YOUNG, S.C. KHOR AND S.K. CHIN.

TABLE 9
Vitamin content of processed snack foods (per 100 g edible portion)

Food Retinol Carotene Vitamin A Thiamine Riboflavin Niacin Ascorbic

Activity Acid
Hg ng Hg RE1 mg mg mg mg

Chocolate products:
Bola coklat berbiskut 8 0 8 0.06 0.39 1.4 3.8
Chocolate wafer (brand 1) 48 0 48 0.06 0.23 1.1 8.8
Chocolate wafer (brand 2) 66 0 66 0.03 0.15 3.6 4.6
Milk chocolate beans 15 0 15 0 0.09 1.1 4.7
Milk chocolate peanuts 57 0 57 0.33 0.40 6.4 1.2
Cereal and tuber products:
Corn stick, chocolate flavour 54 32 59 0.01 0.10 1.4 3.5
Noodle snack, chicken flavour 66 4 67 0.07 0.05 0.8 16.8
Pop Corn, durian flavour 30 146 54 0.05 0.17 0.7 2.9
Snacks, cheese flavour 74 117 94 0.06 0.25 0 3.9
Snacks, chicken flavour (brand 1) 49 50 57 0 0 0.5 1.7
Snacks, chicken flavour (brand 2) 54 43 61 0 0 0.9 2.3
Snacks, chicken flavour (brand 3) 76 32 81 0 0.13 1.9 0.8
Potato chips (brand 1) 160 0 160 0 0.20 4.2 8.5
Potato chips (brand 2) 48 0 48 0 0.06 3.0 15.2
Fish and prawn products:
Fish 'satay' 42 0 42 0.02 0.10 2.0 1.2
Prawn crackers 58 0 58 0.03 0.05 1.0 7.0
Cuttlefish crackers 34 0 34 0.01 0.05 0.8 8.7
Prepared cuttlefish (brand 1) 44 9 46 0.03 0.08 2.9 1.2
Prepared cuttlefish (brand 2) 107 0 107 0.02 0.29 7.0 1.2

'RE = retinol equivalent

24 PERTANIKA VOL. 12 NO. 1, 1989

 NUTRIENT COMPOSITION OF SELECTED COOKED AND PROCESSED SNACK FOODS

TABLE 10
Selected nutrients in each packet of processed snack food

Food Weight Energy Protein Calcium Iron Vitamin A Thiamine Ri bo-

activity flavin
g Real g mg mg ugRE 1 mg mg

Chocolate products:
Bola coklat berbiskut 15 77 1.5 25 0.3 1 0.01 0.06
Chocolate wafer (brand 1) 14 72 0.6 13 0.5 7 0.01 0.03
Chocolate wafer (brand 2) 14 79 0.8 22 0.2 9 0 0.02
Milk chocolate beans 12 45 0.3 6 1.1 2 0 0.01
Milk chocolate peanuts 80 407 13.8 121 2.0 46 0.26 0.32
Cereal and tuber products:
Corn stick, chocolate flavour 15 79 0.8 4 0.2 9 0 0.02
Noodle snack, chicken flavour 30 137 3.3 6 0.3 20 0.02 0.02
Pop corn, durian flavour SO 114 2.1 6 0.5 16 0.02 0.05
Snacks, cheese flavour 10 50 0.9 11 0.3 9 0.01 0.03
Snacks, chicken flavour (brand 1) 18 91 1.1 3 0.5 10 0 0
Snacks, chicken flavour (brand 2) 8 37 0.6 1 0.1 5 0 0
Snacks, chicken flavour (brand 3) 8 36 0.4 1 0.1 6 0 0.01
Potato chips (brand 1) 18 99 1.2 3 0.3 29 0 0.04
Potato chips (brand 2) 10 41 0.6 3 0.1 5 0 0.01
Fish and prawn products:
Fish 'satay1 2 6 23 0.8 5 0.2 3 0 0.01
Prawn crackers 25 106 1.6 23 0.3 15 0.01 0.01
Cuttlefish crackers 20 86 1.0 9 0.4 7 0 0.01
Prepared cuttlefish (brand 1) 40 141 15.0 21 1.8 18 0.01 0.03
Prepared cuttlefish (brand 2) 4 12 2.5 6 0.3 4 0 0.01

'RE = retinol equivalent

2
per stick

REFERENCES Date). ASEAN Protein Project, National Sub-

AMINAH A., K. ROSEUNA and K. NOR1MAH. 1987.
Intake of Commercially Extruded Snack Foods by
Rural Primary School Children. Proceedings of the
2nd Scientific Conference of the Nutrition Society of
Malaysia, February 28, Kuala Lumpur; pp. 69-73.
MOHAMAD NORDIN A.K. 1983. The Nutritional
Aspect of Fast Food in Malaysia. Paper presented
at the MIFF Seminar on Mass Catering and Fast Food,
February 26, Kuala Lumpur.
TEE E SIONG. 1979. Nutrition of School Children
and Snack Foods Sold in Canteens. Paper pre-
sented at the Seminar on Health, Food and Nutrition,
September 15-10, Pulau Pinang.
TEE, E.S., T.K.W. NG and Y.H. CHONG. 1979.
Cholesterol Content and Fatty Acid Composition
of Some Malaysian Foods. Med. J. Malaysia, 33:
334-341.
TEE E SIONG. 1985. Nutrient Composition of
Malaysian Foods-A Preliminary Table (First Up-
Committee Malaysia: Kuala Lumpur.
TEE, E.S., S. SITI MIZURA, I. KHATIJAH, A. MOHD. NA-
SIR, and M.N. ISMAIL. 1986. Status of Food
Composition Studies in Malaysia. Country Report
Presented at the Workshop for the ASEAN Food Data
Network, 22-25 October, Jakarta.
TEE, E.S., S. SITI MI/A RA, R. KUIADEVAN, S.I. YOUNG,
S.C. KHOR and S.K. CHIN (eds). 1987. Labor-
atory Procedures in Nutrient Analysis of Foods.
Division of Human Nutrition, Institute for Medical
Research: Kuala Lumpur.
WORLD HEALTH ORGANIZATION (WHO). 1967.
Requirements of Vitamin A, Thiamine, Riboflav-
ine and Niacin. WHO Technical Report Series
No. 362. WHO: Geneva.
ZANARlAH J. 1986. Junk food - Is the Concept
Right? Teknologi Makanan, 5: 43-48.
(Received 10 November, 1987)

PERTANIKA VOL. 12 NO. 1, 1989 25

Pertanika 12(1), 27-32 (1989)

Dietary Self-Selection by Laying Hens

Offered Choices of Feed
RAMLAH HAMID, R.L HUTAGALUNG and P.N. VOHRA1
Department of Animal Sciences,
Faculty of Veterinary Medicine and Animal Sciences,
Universiti Pertanian Malaysia,
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

Key words: Self - selection; choice-feeding; laying hen

ABSTRAK
Satu kajian pemberian makanan jangka pendek telah dijalankan dengan ayam penelur dengan tiga per-
lakuan (T1-T3). Tiga perlakuan makanan tersebut ialah makanan ayam penelur komersial (Tl, satu-
pilihan), makanan ayam penelur komersial dan grit cengkerang dram (T2, dua-pilihan) dan campuran
jagung, campuran mil kacang soya dan grit cengkerang tiram (T3, tiga-pilihan) dengan menggunakan 20
ekar ayam untuk satu perlakuan. Kajian inijuga melibatkan pemerhatian ke atas tiga jangka masa pemberian
makanan iaitu dari 0800-1200 tgh., 1200-1600 ptg. dan 1600-0800 pagi. Makanan dan air minuman
disediakan secara ad libitum. Ayam diberikan 15 jam cahaya. Pengambilan makanan, tenaga dan protein
adalah ketara tingginya untuk makanan ayam komersial dan adalah rendah bagi makanan dua perlakuan
yang lain. Pengambilan kalsium adalah ketara rendahnya bagi perlakuan satu-pilihan dan tiga-pilihan. Bagi
semua perlakuan, ayam lebih memakan makanan pada waktu jangka masa 1600-0800 pagi (40.8-45.8%
makanan dimakan), iaitu pada waktu leivat petang dan awal pagi. Pengambilan makanan adalah lebih
kurang sama (26.2-29.7% makanan dimakan) bagi jangka masa antara 0800-1200 tgh. dengan 1200-
1600 ptg. Pengeluaran telur, berat keseluruhan telur dan tukaran makanan adalah ketara baiknya untuk
perlakuan dua-pilihan dan tiga-pilihan dibandingkan dengan perlakuan satu-pilihan. Tidak ada perbezaan
daripada segi berat telur di antara ketiga-tiga perlakuan.

ABSTRACT
A short term dietary self-selection experiment (28 days) was conducted with laying hens fed three treatments
(T1-T3) consisting of a commercially available layer feed (Tl, single choice), commercially available layer feed
and oyster shell grit offered separately (T2, two choices) and a corn mixture, soyabean meal mixture and oyster
shell gut offered separately (T3, three choices). Measurement of feed intake was divided into three feeding periods,
from 0800-1'2(X)h, 1200-16Q0h and 160(W800h. Feed and xvater were available zA libitum. Hens were raised
under 15 hour photoperiod. Feed, energy and protein intakes were significantly greater on the commercial diet
and lower in the other two treatments. Calcium intake was significantly lower on the single choice and three-
choice diets. In all treatments, the hens tended to consume more feed during the period between 1600-0800h
(40.8-45,8% feed consumed), that is very late in the evening and very early in the morning. Similar amounts
of feed (26.2-29.7% of feed consumed) were consumed between 0800-1200h and 1200-1600h. Egg production,
egg mass and feed conversion were significantly better on the two and three-choice diets as compared to the
single diet. There was no difference in egg weight among the three treatments.

INTRODUCTION nutrients. Dietary selection by laying hens maybe

Hens have the ability to discriminate between involved with the bird's requirement for differ-
feeds on the basis of different contents of major ent nutrients at specific times of the day

'Department of Avian Sciences, University of California, Davis, California 95616.

 RAMLAH HAMID, R.I. HUTAGALUNG AND P.N. VOHRA.

(Emmans 1977). It is widely accepted that lay- days. Feed intake measurements were divided
ing hens consume energy in excess of daily into 3 periods per day and measurements were
metabolic requirements when offered high carried out at 0800h, 1200h and 1600h for two
energy complete diets (Morris 1968; Summers periods of 7 consecutive days, that is, on days
and Leeson 1976). This overconsumption, 8 to 14 and days 22 to 28, in order to study the
maybe due, in part to the hen's requirement daily pattern of feed intake throughout the day.
for specific nutrients according to the stage of Hens were adapted to the diets 14 days before
egg formation. Part of this pattern is due to the measurements were made.
specific appetite for calcium related to the Egg production and egg weight were
period of shell calcification (Hughes 1972; measured daily. All data were subjected to an
Mongin and Sauveur 1974). analysis of variance and differences between
If all hens in a laying flock could be offered means were determined using the Least Signifi-
a feed which just met their nutrient require- cant Test (Steel and Torrie 1980).
ments, there would be a considerable saving in
both nutrients and cost (Emmans 1978; TABLE 1
Hughes 1984; Cowan (et ai) 1978; Leeson Details of treatments and feed ingredients offered
free-choice to laying hens.
and Summers 1978, 1979). An effective sys-
tem of dietary self-selection is more likely to be Dietary treatment Source of feed
found with a clearer understanding of the way
in which hens react to being given access to two Tl-Single feed Commercial feed.
or more feeds, and the feeding period when T2-Two-choice diet Commercial feed
there is a higher intake of feed. Therefore, a Oyster shell grits.
short term feeding trial (28 days) was conducted T3-Three-choice diet Corn
to study the nutrient intakes and performance Soybean meal
Oyster shell grits.
of laying hens when offered choices of feeds.

MATERIALS AND METHODS TABLE 2

Sixty hens of medium body weight, brown egg Composition of three-choice diet
laying strain (Dekalb Amber Link), 40 weeks of
age were randomly allocated to three treatments Corn mixture Soybean meal mixture
(T1-T3). The treatments (T1-T3) and feed Ingredient % Ingredient %
ingredients offered free choice are given in
Table 1. A commercially prepared diet which Corn 97.95 Soybean meal 98.25
was offered to hens in treatments 1 and 2 had KH,PO, 2.00 KH,PO4 1.50
the following nutrient composition: 11.9 MJ/ Vitamin and 0.05 Vitamin and 0.05
mineral mix. mineral mix
kg metabolizable energy (ME), 16% crude 0.20
DL-Methionine
protein and 3.25% calcium. The composition
and calculated analysis of nutrients in dirts for 100.0 100.0
treatment 3 is given in Table 2. The commer-
cial feed and also the corn mixture and soy- Calculated Composition:
Corn mixture Soybean
bean meal mixture, which are sources of energy meal mixture
and protein, respectively, were offered in mash
form. Feed and water were available ad libitum. Metabolizable energy 13.90 9.95
All the hens were provided with individual feed (MJ/kg)
Crude protein (%) 8.72 41.30
containers. The hens were housed in individual Calcium (%) 0.196 0
battery cages in an open-sided, covered shed.
The hens were given 15 hours photoperiod and
3 hours supplementary lighting was provided at RESULTS
dusk (1830h). Feed intakes of laying hens at the three feeding
Feed and oyster grit intakes were deter- periods over the 28-day experimental period
mined every seven days over a period of 28 are given in Table 3. There was a significant

28 PERTANIKA VOL. 12 NO. 1, 1989

 DIETARY SELF-SELECTION BY LAYING HENS OFFERED CHOICES OF FEED

difference (P<0.05) in total intake among the
three treatments and the three feeding periods
for all treatments. A greater percentage of feed
(40.8 - 45.8%) was consumed in the evening
(after 1600h) and early in the morning (before
0800h) for the three treatments, while percent-
ages of intakes were almost the same (26.2-
29.7%) during the period between 0800h to
1600h for all feeds except for soybean meal
mixture where the least intake (19.4%) was
between 0800-1200h. In terms of oyster shell
grit intakes in treatments 2 and 3 respectively,
the same pattern was observed whereby a higher
percentage (48.4 and 54.1%) was consumed
between 1600h to 0800h and there was no
significant difference between the periods of
0800-1200h and 1200-1600h.
Table 4 shows the calculated nutrient
intakes by laying hens of metabolizable energy,
protein, and calcium during the three periods
of feeding for the three treatments. There were
significant differences in nutrient intakes among
the three treatments and also among the three
feeding periods for each treatment. A higher
ME intake was observed in the 1600-0800h
TABLE 3
Pattern of feed intake (g) of laying hens during the three
feeding periods as means for a 28-day period1
feeding period as compared to the 0800-1200h
and 1200-1600h. Similarly the same pattern of
intakes was observed for protein and calcium
for all treatments.
A summary of total nutrient intake and
production characteristics of hens is shown in
Table 5. Hens on treatment 1 with only a single
feed offered had the highest intakes of feed,
ME and protein. Hens with two-choices and
three-choices as feeding treatments consumed
about the same amount of ME but hens in
Treatment 2 had higher intakes of feed and
protein as compared to hens on Treatment 3.
The highest calcium intake was observed for
hens with two-choices (5.67g) followed by hens
with single-choice (3.75g) and three-choices
(3.65g). There was no difference in egg pro-
duction between treatments 2 and 3 but hens
on treatment 1 had the lowest production. There
was no significant difference in egg weight
among the three treatments.
DISCUSSION
Hens offered feed and oyster grits or corn,
soybean meal and oyster grits separately had

Feeding period
Feed Total
% of Intake 1600- 0800- 1200- SE Intake
0800h 1200h 1600h

Treatment 1:
Feed (g) 52.8a 32.3l 30.3c 2.74 115.4
% Intake/day 45.8 28.0 26.2 100.0
Treatment 2:
Feed (g) 4L2 b 21.T 29.01 1.79 97.9
% Intake/day 42.1 28.3 29.6 94.1
Oyster grits (g) 3.3a 1.5* 1.6"1 0.66 6.1
% Intake/day 54.1 19.7 26.2 5.9
Treatment 3:
Corn (g) 28.5a 20.6b 20.7b 1.54 69.8
% Intake/day 40.8 29.5 29.7 70.6
Soybean meal (g) 8.1a 3.8b 7.7a 0.78 19.6
% Intake/day 41.3 19.4 39.3 19.8
Oyster grits (g) 4.6* 2.5» 2.4h 0.29 9.5
% Intake/day 48.4 26.3 25.3 9.6

*dValues in a row and column (of same variable) with different superscripts are significantly different (P<0.05).
'Values are means of 20 hens.

PERTANIKA VOL. 12 NO. 1, 1989 29

 RAMLAH HAMID, R.I. HUTAGALUNG AND P.N. VOHRA.

TABLE 4
Daily intakes of energy, protein and calcium of laying hens during the
three feeding periods as means for a 28-day period 1

Feeding Period
Total
Intakes 1600- 0800- 1200- SE Intake
0800h 1200h 1600h

Treatment I:
ME (kj) 628 a 384C 361C 32.6 1373
Protein (g) 8.45a 5.17d 4.85 d 0.43 18.5
Calcium (g) 1.72b 1.05c 0.98 d 0.09 3.75
Treatment 2:
ME (kj) 490 b 330 d 345 d 21.3 1165
Protein (g) 6.60 b 4.43* 4.64* 0.29 15.7
Calcium (g) 2.56a t'M€ 1.53lK 0.24 5.43
Treatment 3:
ME (kj) 477 b 324 d ' 365C 14.0 1166
Protein (g) 5.84< 3.37e 4.99d 0.03 14.2
Calcium (g) 1.75' 0.971 0.93 dr 0.11 3.65

''Values in a row and column (of same variable) with different superscripts are significantly different (P<0.05).
'Values are means of 20 hens.

TABLE 5
Nutrient intakes and production characteristics of laying
hens as means for a 28-day period1

Tl T2 T3
Parameters Single Two Three SE
choice choices choices

Feed intake (g) 115.4' 104.01' 98.9C 1.8

Commercial feed (g) 115.41 97.9b _
Corn (g) - - 69.8
Soybean meal (g) - - 19.6
Oyster grits (g) - 6.1 9.5
ME intake (kj) 1373* 1165b 1166 b 18
Protein intake (g) 18.5a 15.7" 14.2C 1.2
Calcium intake (g) 3.75 b 5.43a 3.65' 0.8
Egg production (%) 64.6 b 74.4 a 75.8 a 1.5
Egg weight (g) 57.2 a 57.4* 57.8* 0.8
Egg mass (g) 37.0 b 42.7 a 43.8 a 0.9
Feed conversion 3.12C 2.44 b 2.26a 0.1
ratio (kg feed/ kg egg)
Feed efficiency 0.32r 0.41 b 0.44* 0.05
( k g e g g A g feed)
af
Values in a row with different superscripts are significantly different (P<0.05).
'Values are means of 20 hens.

30 PERTANIKA VOL. 12 NO. 1, 1989

 DIETARY SELF-SELECTION BY LAYING HENS OFFERED CHOICES OF FEED
lower feed intakes than hens fed single diet,
which was a reduction of about 10% and 14,3%
respectively as compared to the control offered
single choice. This reduction in feed intake was
greater than that reported by Leeson and
Summers (1978) which was only about 7%. The
reduction in feed intake was associated with
reduction in energy and protein consumed and
an increase in calcium intake, especially for
hens on treatment 2. Evidence of an increased
feed intake in an apparent attempt to maintain
a higher calcium intake has also been reported
by Classen and Scott (1982).
The feed intake, ME and protein intakes
of the hens for all treatments was greatest in the
late evening (1600-2100h) and early morning
(0630-0800h). This may have been related to
increased requirement for nutrients for egg
formation, as it was observed that 48% of the
eggs were laid before lOOOh, 38% between 1400-
1600h and 14% after 1600h. This is similar to
the findings by Holcombe et al (1976) where
increased intakes of nutrients especially pro-
tein were recorded late in the afternoon. The
oyster shell grit intake was increased during
afternoon and late evening, that is during the
feeding period of 1200-1600h and 1600-080()h,
which relates to the high calcium requirement
for synthesizing shell late at night. This
observation is similar to the finding of Smith
et al (1972) and Hughes (1972). The high
calcium intake in treatment 2, which was about
45% higher than that for treatment 1 was
associated with an increase in egg production
by 15%. Therefore, a separate feeding of a
calcium source influenced egg production as
can be seen in treatments 2 and 3 and this
would depend on the composition of the
complete diet.
Although hens fed on two or three choice
diets consumed less feed, they produced a
greater egg mass, due to the higher number of
eggs produced. This finding is similar to that of
Leeson and Summers (1979). Hens practising
diet selection also utilised feed more efficiently
(0.41 and 0.44) as compared to hens con-
suming commercial diet only (0.32). This is in
agree men I with the findings of Leeson and
Summers (1979).
In conclusion, the study shows that the
ability of hens to regulate their diet is not well
PERTANIKAVOL. 12 NO. 1, 1989
defined. When offered calcium separately,hens
tend to select a higher calcium intake even
when the food offered is high in calcium. When
they were offered separate sources of energy,
protein and calcium, hens were able to select
calcium to meet their requirement with mini-
mum protein and ME intakes to support high
egg production. In this study, hens also con-
sumed more feed during the early hours in the
morning and late in the evening. This is attrib-
uted to their high requirement for energy and
protein for egg formation and high calcium
intake for calcification and also to the cooler
environment during these periods.
REFERENCES
CLASSEN, H.L. andT.A. SCOTT. 1982. Self-selection
of Calcium during the Rearing and Early Laying
Periods of White Leghorn Pullets. PoulL Sci. 61:
2065-2074.
COWAN, P.J., W. MlCHIE and DJ. ROELE. 1978.
Choice Feeding of the Egg Type Pullet. Br. PoulL
ScL 19: 153-157.
EMMANS, G.C. 1977. The Nutrient Intake of Lay-
ing Hens Given a Choice of Diets in Relation to
Their Production Requirements. Br. PoulL ScL 18:
227-236.
EMMANS, G.C. 1978. Free-choice Feeding of leav-
ing Pullets. In Recent Advances in Animal Nutrition,
W. Haresign and D. Lewis (eds) pp. 31-39.
HOLCOMBE, D.J., D.A. ROLAND and R.H.
HARMS 1976. The Ability of Hens to Regulate
Protein Intake when Offered a Choice of Diets
Containing Different Levels of Protein. PoulL ScL
55: 1731-1737.
HUGHKS, B.O. 1972. A Circadian Rhythm of Cal-
cium Intake in the Domestic Fowl. Br. PoulL Sci.
13: 485-493.
HUGHKS, B.O. 1984. The Principles Underlying
Choice Feeding Behaviour in Fowls - with Special
Reference to Production Experiments. World's
Poultry ScL Journal 40 (2): 141-150.
LEESON, S. andJ.D. SUMMERS. 1978. Voluntary
Food Restriction by Laying Hens Mediated
Through Dietary Self-selection. Br. PoulL Sci. 19:
417-424.
LKSSON, S. and J.D. SUMMERS. 1979. Dietary Self-
Selection by Layers. PoulL Sci. 58: 646-651.
LESSON, S. and J.D. SUMMERS. 1980. Effect of
Early Light Treatment and Diet Self-selection on
leaving Performance. PoulL Sci. 59: 11-15.
31
 RAMIAH HAMID, R.I. HUTAGALUNG AND P.M. VOHRA.

MONGIN, P. and B. SAUVEUR. 1974. Voluntary Food STEEL, R.G.D. and J.H. TORRIE. 1980. Principles and
and Calcium Intake by the Laying Hen. Br. Poult. Procedures of Statistics. 2nd. Ed. New York: McGraw
Sci. 15: 349-360. Hill Book Co.
MORRIS, T.R. 1968. The effect of Dietary Energy SUMMERS, J.D. and S. LEESON. 1976. Energy and
Level on the Voluntary Calorie Intake of Laying Protein Requirements of Laying Hens. Nutr. Rep.
Hens. Br.PouU. Sd. 9: 285-295. Int. 17: 87-91.
SMITH, W.K., P.D. RALLARD, and H.V. BIELLIER. 1972.
The Calcium Appetite of the Hen. Proceedings of
the Australasian Poultry Science Convention, pp. 49-
54. (Received 29 September, 1987)

32 PERTANIKA VOL. 12 NO. 1, 1989

Pertanika 12(1), 33-41 (1989)

Determination of Meat Content in Processed

Meats Using Currently Available Methods
A.S. BABJI, P.H. OOI and A. ABDULLAH
Department of Food Sciences and Nutrition
Universiti Kebangsaan Malaysia
43600 Bangi, Selangor D.E., Malaysia

Key words: Meat content; methods; processed meats.

ABSTRAK
Empat kaedah digunakan dalam penentuan kandungan daging dalam produk daging terproses. Kaedahyang
dirujukkan sekarang tidak begitu tepat dan payah diguna untuk pemonitoran mutu kandungan daging. Satu
kaedah yang tefmt dan praktikal diperlukan untuk menentukan kandungan daging dalam produk daging
terproses. Kaedah analisis pigmen jumlah dan mioglobin didapati lebih sesuai untuk penentuan kandungan
daging pada burger lemlru tempatan, jika dibandingkan dengan kaedah Pearson (1975) atau Stubbs &
More (1919).

ABSTRACT
Four methods were used in the determination of meat content in local meat and meat products. Current methods
available are not reliable and applicable to routine monitoring and quality control by the regulating laboratories
as well as the meal processing industry. A reliable and practical method is needed to monitor meat products
and ensure that they are meeting the minimum requirement of sixty five (65 %) percent meat content. The
total pigments and myoglobin technique was found to be applicable for determination of meat content in
locally processed beef burgers, when compared to the Modified Method of Stubbs & More (1919), and Pearson
Method (1975).

INTRODUCTION (1981) described burger to contain 80% meat

There has been a significant increase in the
production of value added meat products and
their selection in the chilled and frozen sec-
tions of major supermarkets in Malaysia. In 1983
the value of processed food originating from
livestock exceeded 1,600 million Malaysian
Ringgit. Malaysia is expected to import about
426,902 beef cattle to provide for 32,000 ton of
beef by the year 1990 (Idrus 1981). Currently,
the country uses about 114,500 ton per year,
valued at about 300 million Ringgit (Malaysian
Business 1980). Processed meats, especially
hamburgers and frankfurters are common food
items to many people in Malaysia as can be seen
by the many fast food chains and local burger
stands. In the U.S.A., a hamburger is defined
as chopped/minced meat with or without added
fat and spices (de Holl 1976). It should also
contain not more than 30% fat. Wilson et al
with the rest comprising cereal, water binding
materials, flavours and spices. In Malaysia, there
is no clear definition and specifications for
burgers. Manufacturers in their efforts to cut
cost, often use meat substitutes such as cereals,
soya proteins, groundnuts and lately mechani-
cal deboned meat to formulate the hamburgers
and frankfurters. The problem is: how does
one goes about quantifying meat content in
such products, so that consumers can be as-
sured that such products contain a minimum
of 65% meat as stipulated by the Food Law and
Regulation (1985). A recent development in
the import of beef should also be mentioned
in this content. Most beef was imported from
Australia until 1983, when the government
decided to open its meat market to India. Today,
due to its lower price, much of the beef (more
than 90%) is imported from India, usually from
 AS. BABJI, P,H. OOI AND A. ABDULLAH
the fore quarters. It is cheaper than even the
imported soya isolate and concentrate, which
would lead one to think that manufacturers
would use more meat (at least 65%) so as not
to contravene the food regulation. But this is
not so. Instead, manufacturers go for formula-
tions consisting of Indian beef (40-60%), soya
proteins (10-80%), wheat/tapioca flours, me-
chanically deboned meat and egg proteins to
come up with the least cost. Soya protein is
popular because of its high waterholding ca-
pacity, good texture, and bulkiness in weight
when hydrated. Therefore, in today's beef
burger, the two major components are Indian
beef and soya protein concentrate. This paper
is aimed at discussing currently available meth-
ods for quantifying meat in such products and
their usefulness in routine quality control checks
for the minimum requirement of 65% meat
content for regulating purposes.
MATERIALS AND METHODS
Local beef type D (fresh cut) from class 'Bull'
or Cow were purchased from a local market.
Indian Beef (deboned) type F, forequarter was
purchased from a local meat processing com-
pany. Local beefburgers (12 brandnames) were
obtained from the supermarkets. All meat
samples were completely homogenized, stored
in airtight bottles in a freezer at -20°C until
ready for analysis.
Standard Beef-Soya Cereal Blend
Beef soya protein cereal blends were prepared
to test the accuracy and percent recovery using
methods currently available. The seven blends
prepared were as follow:
Indian Beef (80:20) SPC (1:2) Wheat flour
100 0 0
A
B 90 5 5
C 80 15 5
D 70 25 5
E 60 35 5
F 50 45 5 where
G 40 55 5
Total protein from such blends would include
meat protein, soya protein and gluten protein.
For calculation of meat protein, a correction
formula as suggested by Pearson (1975) was
used.
34 PERTANIKA VOL. 12 NO. 1, 1989
Meat protein = Total Nitrogen - (KaC +
x 6.25 where
Ka = conversion factor of cereal to nitrogen
C - total cereal measured
Kb = conversion factor of soya proteins to
nitrogen
S = total soya measured.
In this study a 70% protein content soya con-
centrate was hydrated with 2 parts water mak-
ing a Kb value equal to 0.7 x 1/3 x 1/5.71 =
0.0409. Ka is given a value of 0.02 x 0.69
with the assumption of cereal (wheat) contain-
ing 69% carbohydrate and 2.0% nitrogen
(Pearson 1975).
Determination of Total Pigment in Meat,
Burger and Blend Standards (Rickansrud &
Henrickson, 1967)
A 25g sample was blended in 100ml distilled
water for 3 minutes. The homogenate was
centrifuged at 2000 x g at 6°C for 15 minutes
using the MSE Coolspin Centrifuge. The super-
natant was filtered through 3 whatman filter
paper. The precipitate was mixed with another
100ml distilled water, centrifuged again and
filtered to finally obtain a 200ml of homo-
genate solution.
A 20ml aliquot was pipetted into a 50ml
Erlenmeyer flask and 4mg potossium ferricya-
nide added. Total pigment concentration in
m
g / g ( w e t weight) was calculated from absor-
bance reading at 540nm using a Bausch and
Lomb Spectronic 20. For fresh meat samples
calculations were made based on wet weight,
dry weight, wet weight fat free basis and
dryweight fat free basis.
Calculation for total pigment was obtained
using the formula:
Cone total pigment in mg/g wet weight
= Absorbance x K
wet weight sample
K - 17,000 x aliquot vol (L) x dilution factor
and E = Coefficient extinction myoglobin
= 11.3 nM/L
 DETERMINATION OF MEAT CONTENT IN PROCESSED MEATS USING CURRENTLY AVAILABLE METHODS
Determination of Myoglobin Content Using
the Poel-Cyano Method (Topel, 1949)
The pH of meat samples was determined using
the AOAC method (1980). A lOg sample was
homogenized for 2 minutes in cold water mixed
with X ml INHySO4 in a waring blender.
X * (pH sample -5) x 0.35
The homogenate was centrifuge at 3000 rpm
for 2 minutes in a polyethylene tube (50ml)
using the MSE Desk centrifuge. The supernatant
obtained was transferred to a 50ml tube and
heated slowly to reach a temperature of 54°C
after which it was soaked in a water bath to
reach 25°C. The homogenate was placed in a
100ml beaker and the pH brought to 7.2 using
Na2 CO r The homogenate was transferred back
to a 50ml tube and centrifuged for 10 minutes
at 2500 rpm. The supernatant was filtered into
a 50ml Erlenmeyer flask and 2-3 small crystals
of potassium ferricyanide added. Absorbance
was read at 540 nm using the Spectronic 20.
Calculation of myoglobin (Mb) derived by Poel-
Cyano (Topel, 1949):
mg Mb/g wet tissue = absorbance x 7.50
Results were expressed in mg/g wet weight for
burger samples and the blend standards, while
for fresh meat samples, calculations were made
on dry weight, dry weight fat free basis and wet
weight fat free basis.
Methods Used for Quantitation of Meat Content
Using Myoglobin Content as an Index The Poel-
Cyano (Topel 1949) described above was used
to quantitate meat content. Calculation was
made on a fat free dry weight basis to avoid the
wide variation in fat distribution as well as added
fat in the samples.
Calculation:
Fat free meat = myoglobin content (mg/g)
where f} = conversion factor myoglobin
in meat, i.e. amount of myo-
globin in 100% Indian beef
calculated in mg/g fat free
weight basis
= 4.93 mg/g
PERTANIKA VOL. 12 NO. 1, 1989
Total meat content meat (fat free) and fat
content.
The conversion factor for myoglobin to
fat free meat (fl) is obtained from our analysis
of myoglobin content in Indian beef. This is
because most meat producers use Indian beef
as the meat component.
Using Total Pigment Content as an Index The
total pigment content was determined from an
analysis of the meat commonly used for beef
burger processing, i.e. Indian beef.
Total pigments was expressed in mg/g for
free basis to avoid high variation from fat present
or added to beef burgers.
Calculation:
Total Pigments of fat free meat
= total pigments content (mg/g wet wt)
f2
where f2 = conversion factor of total pig-
ments in meat i.e. amount of
total pigment in 100% Indian
beef calculated in mg/g fat free
weight basis
= 6.31mg/g
Total meat = meat fat free + fat content
Quantitation of Meat Content Using Modified Method
ofStubbs and More (1919) The original formula
by Stubbs and More (1919) does not account
for nitrogen coming from cereal or soya pro-
teins. The conversion factor of nitrogen to meat
is 3.55 for beef (Analytical Committee 1963)
and is reported on fat free basis. Pearson (1975)
stated that correction is necessary for total
nitrogen if there is cereal inclusion. In this
paper, total nitrogen is corrected for cereal and
soya protein.
Assuming cereal is from wheat:
% cereal nitrogen = KaCr
where Cr = cereal content (AOAC, 1980)
in %
Ka = conversion factor to nitrogen
= 0.02 x 0.69 with assumption cereal
contained 69% CH(,0 and cereal
wheat contained 2% nitrogen
(Pearson, 1975)
 A.S. BABJI, P H . OOI AND A ABDULLAH

Assuming soya protein as the major non-meat LM = 100 Nt - FeNm ( 1 0 0 - 1)

- KfC - KaS
protein FT"
Nm (1 -Fl/Ff)
% soya nitrogen = KbS
where S= % soya protein cone. (70%) protein
where S = Soya protein cone. (70% protein,
hydrated 1:2 (SPC: H,O)
1:2 soya:H2O (H2O)
Ka = % nitrogen in soya protein cone.
Kb = conversion factor to nitrogen
= 0.70 x 1/3 x 1/5.71
= 0.70 x 1/3 x 1/5.71
= 0.0409 Statistical Analysis
Analysis of Variance (ANOVA), F Values, and
.*. meat nitrogen (fat free) DMR tests were used in data processing, using
= 100 N t - KaCr- KbS the Statistical Analysis System (SAS).
F RESULTS AND DISCUSSION
where, A proximate analysis was done on various meat
Nt = total nitrogen cuts commonly used in the meat industry. For
comparison purposes, we have included heart
KaCr = nitrogen from cereal tissue, sirloin cut, Indian, beef (imported) and
Kbs = nitrogen from soya also mechanical deboned beef (MDB). Table 1
F - %fat shows the major components from various meat
cuts and sources. Protein content ranged from
Total meat = meat (fat free) + fat 20-30%, with MDB at 4.5% protein. The fat
Quantitation of Fat Free Meat Using Modified Pearson content ranged from 1.4-8.4% with MDB at
Method (1975) In some countries meat (fat 16.8%, Water content for most cuts is about
free) is used to indicate meat content in meat 75% and ash varies from 0.83-1.34%. Indian
products. Pearson (1975) put forward some beef contained very little fat (1.4%), a factor,
formulas for calculating fat-free meat, the most which should be noted for several reasons later.
common being: The composition of MDB is also different than
meat (fat free) = lOONt-FeNm (100-1)- (KfC) normal meat cuts, fat (16%), protein, (4.5%)
Ff and ash at 2.80%. Its utilisation in locally proc-
Nm (1 - Fl/Ff) essed meat is increasing and therefore should
be monitored.
where The wide variation in meat components
must be considered in methodologies for meat
Nt = % total nitrogen in product
content determinatin. Pearson (1975) in his
Fe = % extracted fat
formulation for lean meat suggested fat level
Nm = 3.55, conversion factor nitrogen to not exceeding 10%, when determining lean
beef meat. Muscle fat has been reported to range
Ff = % extraneous fat max 90% from 1-18%. In this study the fat content ranged
Fl - % intramuscular fat max 10% from 1.4-8.4%.
Kf = % nitrogen in cereal, calculated on
Concentration of Pigment in Meat
dry CH2O basis
Total pigment and myoglobin contents were
C = % carbohydrate in product determined in local meat cuts commonly used
An adjustment is needed in the above in processing. Although myoglobin comprises
formula to take into account the nitrogen three fractions, with different absorptions,
contribution from soya protein in a beef-soya- Rickansrud & Henrickson (1967) noted that
cereal added beef burgers. these differences can be obstructed by convert-
ing them into the cyanmetmyoglobin form.
Therefore
Table 2 shows the concentration of myoglobin
meat (fat free) or lean meat, LM; in local meat cuts on wet weight, and dry weight

36 PERTANIKA VOL. 12 NO. 1, 1989

 DETERMINATION OF MEAT CONTENT IN PROCESSED MEATS USINC CURRENTLY AVAILABLE METHODS

TABLE 1 on a normal and fat free basis. Myoglobin

Proximate composition of various meat cuts (lean) content from sirloin cuts of beef and buffaloes
are higher when compared to other cuts.
Meat Protein Fat Moisture Ash
Rickansrud and Henrickson (1967) reported
Cuts1 % % % %
similarly high values in longissimus dorsi muscles.
IB* 20.0bc 1.4de 76.3 a 0.83' f Heart tissue had 6.38 mg/g of myoglobin while
MDB* 4.5d 16.8;t 76.2" 2.86* deboned beef had 6.19 mg/g of myoglobin.
MBC 18.9' 4.1cd 75.0* 1.04cd Indian Beef had 4.86 mg/g while local buffalo
MBT 2L5 ;u 8.4'' 72.5" 1.04cd brisket trimmings and chuck ranged from 2.78-
MBB 20.0')( 3.8 c d 77.3 ;1 1.12' 4.30 mg/g. Indian beef (Buffalo, Type F)
MBS 20.0 b ' 5.1 C 76.5' 0.87 d H seemed darker in colour when compared to
MBFC 21.6"'' 76.2 d ISO*** local beef cuts. The mean value of myoglobin
MBFT 22.7* 4.1cd 74.9" 0.78'
concentration for all cuts is 4.03 mg/g with
MBFB 22.6*1 L9 dt " 74.5 a 1.34 b
MBFS 23.0* 1.5dt" 75.5* 0.90 d H
standard deviation of 1.11 mg/g.
MBH 20.3'" 0.3 c 78.4 a 1.01 ede A detailed study was investigated on the
myoglobin content of Indian Beef. This was in
'Means of 2 samples IB = Indian Beef view of the fact that most manufacturers used
-Means of 6 samples MDB = Mechanical D e b o n e d
Beef
Indian Beef as a major component in burger
*Means of 3 samples MBC = Malaysian Beef, Chuck production. Adjustment of meat on a lean, dry
MBT = Malaysian Beef, Trimming weight basis reduced the coefficient of variation,
MBB = Malaysian Beef, Brisket (Appendix 1) from 23.78% and 23.80% to 20.11
MBS - Malaysian Beef, Sirloin and 20.13% respectively. Thus it is better to
MBFC = Malaysian Buffalo, Chuck
MBFT = Malaysia Buffalo, Trimming
express the content of myoglobin on a dry
MBFS = Malaysian Buffalo, Sirloin weight basis. A total of six samples of Indian
MBH = Heart Muscle beef were analysed to obtain a mean value for
Means with different superscripts within each column
myoglobin concentration that is reliable for use
indicate significant differences (P < 0.05) as a reference. The concentration mean of 4.93
mg/gm on fat free basis wet weight basis was
TABLE 2 chosen, because of the variable fat content in
Myoglobin content of selected meat cuts (lean) various meat cuts and the fact that added fat is
part of the burger formulation in most instances.
Meat WVt D,T Wet Dry
The total pigment content in loal beef
Cuts' weight weight fat free fat free
cuts ranged from 3.90-5.36 mg/g, local buffalo,
mg/g 3.96-5.27 mg/g; Indian Beef, 6.23 mg/g and
2 ;;^,.., beef heart, 7.55 mg/g. (Table 3). Total pig-
IB 4.86" td 20.38* 4.93"
ment includes hemoglobin, myoglobin, cyto-
MDB* 6.19*h 26.02"" 7.44" 88.69"
4.25 t(ilf
chrome, vitamin B12, heme pigment and
MBC 17.32' 'i(' 4.44 h< 22.13' } ( d
MBT 11.78^ 3.53 h ' d 17.02 (<t
flavour. It should be noted that 95% of the iron
MBB 2.78' 12.34 d r 2.90 d 15.15' 1
in a piece of meat is in the myoglobin com-
MBS • 476bcd
20.28 h ' 5.02 h 25.96 1 " ponent (Clyesdale and Francis 1976). Other
MBH 6.38* 29.91* 6.47 ;i 30.19'' factors like genetics, muscle types, and handling
MBFC 4.30 c d e 18.07 ( d 4.40 h l ( i 19.82' H(i during preslaughter could affect the pigment
MBFT 3.40 d e l 13.88<u* 3.34<fi 16.36' fi content in meat. (Rosenman 8c Morrison 1965,
MBFB 3.25' f 12.95 d( - 3.32' d 14.05 d Brown 1962, Livingstone & Brown 1980).
MBFS 4.92 h< 20.081"1 5.00 h 21.40'""
In using total pigment as an indicator of
1 = Means of 2 samples meat content, it was observed that total pig-
2 = Means of 6 samples ment on a fat free dry weight basis resulted in
3 = Means of 3 samples the lowest coefficient of variation, (18.33%,
Means with different superscripts within each column Appendix 1). With total pigment, the conver-
indicate significant differences (P < 0.05). sion factor used to obtain meat content was
PERTANIKAVOL. 12 NO. 1, 1989
 AS. BABJI, RH. OOI AND A. ABDULLAH

based on fat free wet weight basis of 6.31 m g / mix-ture of meat soy patties was formulated
to test recovery and accuracy of the m e t h o d s
used.
TABLE 3
Table 5 shows the recovery of m e a t from
Total pigments content in selected meat cuts
(lean) beef-soy mix standard, using myoglobin, total
pigments a n d Mg as indicators. T h e recovery
Meat Wet Dry Wet Dry rate (means and standard deviations) are 97.6
Cuts weight weight weight weight ± 6.2, 96.6 ± 6.87 a n d 107.4 ± 12.5% respec-
fat free fat free tively. T h u s b o t h myoglobin a n d total p i g m e n t s
m m m m
are similar when used as indicators for testing
g/g g/g g/g g/g
recovery of meat content. T h e Mg m e t h o d by
IB 6.23 al)
22.70< lh
6.31" b
27.781" Stubbs a n d More (1919) is m o r e variable, d u e
DNM 6.54 lh 27.50*tb 7.86" 93.78il to Mg being contributed from o t h e r sources
MBC 5.13'" 20.82ab 5.331" 26.47b<
than just meat itself. Table 6 shows t h e Mg,
MBT 4.82'K I7.58b 5.26'" 25.38b<
myoglobin a n d total p i g m e n t c o n t e n t s in local
MBB 3.90' 17.33h 4.06c 21.26f
5.54'M beefburgers. T h e values of meat recovered using
MBS 5.26b< 22.38 ab 28.66'"
MBH 7.55a 32.36" 7.58i( 35.47b the pigments, Mg a n d a modified Pearson
MBFC 4.92b< 20.67ilb 5.02bt 22.68'" (1976) m e t h o d s for meat quantification in local
MBFT 3.96< 16.30l) 4.14* 19.16* b e e f b u r g e r s is shown in Table 7. T h e c o n t e n t s
MBFB 4.12( 16.45" 4.20< 17.87* of meat ranged from 22.3%-65% (myoglobin as
MBFS 5.27b< 21.50* b 5.35 bc 22.90'" index) 23.5-71.1% (total pigments as i n d e x ) ;
26.7-71.7% (Mg as index)" a n d 12.3-53.9%
Means with different superscripts within each column
(with modified Pearson m e t h o d ) . From these
indicate significant differences (P < 0.05).
m e t h o d s , it can be c o n c l u d e d that myoglobin
The determination of total nitrogen is not a n d total pigments are reliable indicators to use
able to give an indication of how much meat for quantification of meat in mix p r o d u c t s
is in the mixed products such as hamburgers. such as beef burgers a n d frankfurters.
This is shown in Table 4. However, data from Stubbs a n d More's (1919) m e t h o d using
Table 4 and Figures 1 and 2 showed a strong magnesium is n o t reliable because of the pre-
correlation between meat content and myog- sence of the e l e m e n t in soy protein a n d spices,
lobin and total pigment contents. which are commonly used in m e a t p r o d u c t s
Determination of Meat Content in Soy-Beef formulation these days. T h e weaknesses of the
Standard and Local Beef Burgers. modified Pearson Method (1975) which is based
T h r e e m e t h o d s were described earlier in the on Stabbs and More, has b e e n discussed ear-
Methods a n d Materials section. A standard lier.

TABLE 4
Total nitrogen and pigment roncentration in standard bed-soya-cereal mix

Beef soya cereal Total Myoglobin Total

Standard Mix1-2 nitrogen pigment
(%) (%) ( m g/g) (mg/g)

A 100-0-0 2.64 3.68 4.72

B 90 - 5 - 5 2.65 3.38 4.28
C 80-15-5 2.54 3.23 4.04
D 70 - 25 - 5 2.67 2.74 3.51
E 60 - 35 - 5 2.71 2.33 2.90
F 50 - 45 - 5 2.73 1.91 2.31
G 40 - 55 - 5 2.72 1.52 1.90

1 Meat is 80% Indian Beef (lean) and 20% fat

2 Textured soya protein

38 PERTANIKA VOL. 12 NO. 1, 1989

 DETERMINATION OF MEAT CONTENT IN PROCESSED MEATS USING CURRENTLY AVAILABLE METHODS

= 0.97724 r * 0.97361
= 0.910x + 4.47 y > 0.928x - 2.61

/
-..

Meat
Oontent
Recovered 75 • /

/
65 /

55
/
/

IS

IS
40 50 50 9d 10*0

Meat Oontent (%)

Meat Cbntent (Standard)

Fig. I : Regression m we showi ng men ( con ten I recoi wredfrom Fig, 2 : Regression curve showing meat content recovered
beef-soya standard mix using myoglobin as an in- from beef-soya mix i(sing total pigme11 ts as indicator,
dicator.

TABLE 5 than total pigments, which contains hemoglo-

Recovery of meat from beef-soya standard bin and which could be contributed from blood.
Although the amount of myoglobin varies with
Standard Meat1 Meat-
different muscle types, for instance, it is higher
A 92.5 92.7 92.0
in heart muscle, the amount is less variable in
B 91.2 90.3 94.1
C 103.5 101.6 105.9
TABLE 6
D 102.0 103.8 117.0
Magnesium, myoglobin and total pigments
E 106.0 103.8 127.5
contents in local beefburgers
F 96.0 91.8 111.4
G 91.8 90.0 104.25 Brandnames1 Magnesium Myoglobin Total
Pigment
Mean,% 97.6 96.6 107.4 (ppm) (mg/g)
Standard
deviation, ±6.2 ±6.87 ±12.5 FIKA 240" 2.03" 3.1(T
Range, % 91.8-106 90.0-105.8 92.0-127.5 ANGUS 35 3*b 1.761" 2.25'
RAMLY 2311' 2.121' 2.81'
1
Using myoglobin as an indicator WISMA BURGER 286r<l 2.03[> 2.72'
2
Using total pigments as an indicator PRINCE 295 b t " 0.39h 0.5h
< Stubbs 8c More (1919) Method. BIFFI S52* h 1.181 1.53'
KB 1.43' 2.10'
AMIRUL 351"h 0.83* 1.26"
CONCLUSION UTAMA 243" 1.80* 2.51"
Methods using myoglobin and total pigments MESTI-BEST 259" 2.03 b 2.83'
can be used to quantitate the meat content in HALFOMAR 345*b« 1.13' 1.50
meat products. Its inherent variability in meat SAIAM - 1.58"' 2.081
UKM _ 3.04" 3.91'
tissue is well defined but its conversion to cyan-
metmyoglobin form in this procedure reduces 1
Means of 3 samples (wet weight)
its heterogenous variability. Myoglobin is only Means with different superscripts within each column
present in meat tissues and is more reliable indicate significant differences (P < 0.05)

PERTANIKA VOL. 12 NO. 1, 1989

 A.S. BABJI, P.H. OOI AND A. ABDULLAH

TABLE 7
Meat and lean meat content of
local beef burgers using various methods

Brand * Meat1 Meat2 Meat1 Lean meat 4

names (;%) (%) (%) (%)

FIKA 63.2ab ±2.4 71.P ± 1.6 55.5C + 3.7 34.9d( ± 4.1

ANGUS 55.5cdt ±3.0 55.4" ± 2.2 26.7' ±5.9 I5.3*h± 6.2
RAMLY 64.4" ± 1.4 65.9b ± 1.8 n.r ± 1.8 53.9;l ±2.0
WISMA BURGER 59.6bc ±0.6 61.5f ± 2 . 3 55.r ±0.5 39.7 rd ± 1.6
PRINCE 22.3« ±0.7 23.5h ±2.0 34.21' ±5.1 19.2f ± 2.9
BIFFI 34.9' ±6.9 35.7* ±0.2 42.4d ± 3.8 31.1* ± 2 . 2
KB 52.3 dr ± 1.9 56.5 dr ± 2.0 55.9C ± 4.8 33.8dt< ± 5.2
AMIRUL 34.8* ±3.6 35.0* ± 1.1 20.8H ± 0.4 15.9*h± 1.9
UTAMA 56.1 cd ± 1.8 59Acd ± 2.3 32.7ef ± 3.0 12.3" ± 4.0
MESTI-BEST 52.8dt" ± 4.1 56.4dt- ± 3.2 48.8< ± 7.9 40.4cd ± 6.6
HALFOMAR 51.6e ± 2.4 52.5' ±2.4 75.9a ± 4.0 49.3 iib ± 1.8
SAIAM 59,6W ±2.8 60.5c ± 0.4 64.9b ± 1.5 38.7td ± 2.5
UKM* 65.0a ±0.3 65.3b ± 1.6 43.3d ± i.5 44.5b( ± 2.0

'Meat content using myoglobin as indicator

1
Meat content using total pigments as indicator
1
Meat content using Stubbs and More method (1919)
1
Meat content using modified Pearson Method (1975)
*' Mean n = 3
I KM* - Formulated Beef Burger

Indian Beef that is commonly used in the Burger beef standards as a reference using myoglobin
Industry. It is felt that the determination of as an indicator.
meat content in local beef burger can be The conversion factor using myoglobin as
achieved satisfactorily by formulating Indian indicator is 4.93, a mean on a fat free basis for

Appendix 1 Appendix II
Coefficient variation and standard deviation Coefficient variation and standard
of Indian beef samples deviation of 8 selected meat cuts

Mean1 SD CV Mean1 Cone. SD CV

{ Of \
Cone. ( m g/g) (mg/g) \ /O )

(™g/g) (™g/g) (%)

Total pigment
Total pigment Wet weight 4.95 1.00 20.18
Wet weight 6.23 1.32 21.12 Dry weight 19.92 3.85 19.33
Dry weight 22.72 9.06 39.86 Wet weight
Fat free 5.27 1.30 24.58
Wet weight Dry weight
Fat free 6.31 1,36 21.60 Fat free 31.80 25.29 79.52
Dry weight Myoglobin
Fat free 27.78 5.09 18.33 Wet weight 4.03 1.11 27.54
Myoglobin Dry weight 16.59 4.89 29.48
Wet weight 4.86 1.15 23.78 Wet weight
Dry weight 20.38 4.10 20.11 Fat free 4.28 1.45 33.88
Wet weight Dry weight
Fat free 4.93 1.17 23.80 Fat free 26.86 25.16 93.67
Dry weight
Fat free 21.66 4.36 20.13 Mean, n = 8
2
Selected Meat
Cuts
i u p a n n _ p. IB, MDB, MBC, MBT, MBB,
lVlCd.Il, 11 — D
MBTC, MBFT, f vlBFC

PERTANIKA VOL. 12 NO. 1, 1989

DETERMINATION OF MEAT CONTENT IN PROCESSED MEATS USING CURRENTLY AVAILABLE METHODS
Indian beef. This research concludes that this
conversion factor is suitable as a reference
because other meat cuts with higher pigment
contents are uneconomical for use in the burger
manufacturing industry.
REFERENCES
ANALYTICAL METHODS COMMITTEE. 1963. Nitrogen
Factors for Park. Analyst 88: 422.
AOAC. 1980. Official Methods of Analysis, 13th. Ed.
Assn. of Official Analytical Chemists, Washing-
ton. D.C.
BROWN, W.D. 1962. The Concentration of Myoglobin
and Haemoglobin in Tuna Fish./ Fd. Sri. 27: 26.
DeHoLL,J.C. 1976. Encyclopedia of Labelling Meat
and Poultry Products, 3rd. Ed. Meat Plant Maga-
zine. St. Louis.
[DRUS, A.Z. 1981. Food Production in Malaysia. Paper
presented at Amino Arids Nutrition Seminar, Aji-
nomoto (M) Bhd. 3. Dec 1981.
LIVINGSTONE, DJ. and W.D. BROWN. 1980. The Che-
mistry of Myoglobin and its Reactions. Fd. Tech-
nol, May 1981.
PERTANIKA VOL. 12 NO. 1, 1989
MALAYSIAN BUSINESS. 1980. Berita Publishing, K.L.
PEARSON, D. 1975. The Examination of Meat Pro-
ducts with Special Reference to the Assessment of
Meat Content. Analyst. 100 (1187): 73-81.
RICKANSRUD, D.A. and R.L. HENCRICKSON. 1967. Total
Pigments and Myoglobin Concentration in Four
Bovine Muscles. / Food Sri, 32: 57.
ROSENMANN M. and P. MORRISON. 1965. Seasonal
Angumentation of Myoglobin in Snowhoe Hare.
/. BioL Chem. 240: 3353.
STUBBS, G. and S. MORE. 1919. Estimation of ap-
proximate Quantity of Meat in Sausages and Meat
Poste. Analyst 44: 125.
TOPEL, D.G. 1949. Determination of Myoglobin in
Pork Muscle. Adapted from Poel-Cyano Method.
Ann.]. PhysioL 156: 44-51.
WILSON, N.R.P., EJ. DYETT, R.B. HUGHES, and C.R.V.
JONES. 1981. Meat and Meat Product: Affecting
Quality Control. London: App. Sci. Pub.
(Received 29 October, 1988)
41
Pertanika 12(1), 43-46 (1989)

COMMUNICATION I
Kesan Pendedahan Racun Karbaril terhadap
Tumbesaran dan Pembiakan Aphis craccivora Koch
ABSTRAK
Dalam kajian ini kesan beberapa kepekatan racun karbaril terhadap beberapa parameter jadual hayat Aphis
craccivora Koch telah diselidiki. Telah didapati bahawa karbaril pada kepekatan melebihi 0.0106% b.a.
dapat membunuh krsemua nimfa instar awal Pendedahan pada kepekatan ini pula mengakibatkan kele-
watan pembiakan selama tiga hari pada tiga peratus daripada populasi dewasa yang telah didedahkan dan
umur hidupnya hanya 10 hari, iaitu suatu kesingkatan hayat yang bererti berbanding dengan 29 hari dalam
kawalan. Nilai-nilai parameter jadual hayat juga menunjukkan bahawa populasi aftd yang berkenaan tidak
akan bertambah dan oleh itu tidak mampu melanjutkan hayatnya ke generasi hadapan. Sebaliknya terdapat
8% yang mencapai kedewasaan bagi populasi yang terdedah kepada kepekatan 0.0052% b.a., 40% bagi
kepekatan 0.0027% b.a. dan 68% bagi kepekatan 0.0013% b.a.

ABSTRACT
The effects of several concentrations of carbaryl on the life-table parameters of Aphis craccivora Koch were
studied. It was revealed that all the early nymphal instars died after being exposed to carbayrl at concentra-
tion in excess of 0.0106% a.L Exposures at this concentration however, resulted in a 3-day delay in re-
production in 3% of the tested population and experienced a significantly shorter life span of 10 days when
compared to 29 days in the control. Values for the life-table parameters also indicated that these populations
would not grow and thus would be unable to continue into the next generation. In contrast, 8% of the popu-
lation exposed to 0.0052% a.i. reached adulthood, 40% for those exposed to 0.0027% a A. and 68% for
those at 0.0013% a.i.

PENGENALAN dan Parrella, 1984). Ada juga kes-kes lain yang

Afid atau kutu daun melalui kitaran hidup yang
agak luar biasa dan lazimnya mempunyai kadar
pembiakan yang tinggi. Ianya berkebolehan
melahirkan progeni secara berseks atau tidak
berseks, iaitu partenogenesis. Di dalam iklim
tropika seperti Malaysia ini afid Aphis craccivora
Koch yang merosakkan tanaman kekacang,
kukurbit, dsbnya, membiak dengan cepat
dengan ketiadaan halangan dari faktor cuaca.
Pada suhu serta kelembapan yang tinggi kadar
pembiakan dan kesuburan afid ini adalah tinggi
juga (Radke, et al, 1972).
Selain dari menghisap cairan sap tana-
man, afid ini juga berkemungkinan berfungsi
sebagai vektor penyakit virus pada tanaman
kekacang, cili dan kukurbit. Oleh itu kawalan
kimia secara meluas bukan lagi perkara asing.
Pada kes-kes tertentu racun serangga pada
kepekatan submaut telah merencatkan pem-
biakan serangga dan hamama (Hall, 1979Jones
menunjukkan perangsangan racun serangga ini
menyebabkan kenaikan kadar pembiakan yang
lebih tinggi, iaitu suatu keadaan yang digelar
hormoligosis (Luckey, 1968). Di Tamil Nadu,
India, populasi Aphis gossypii telah didapati ber-
tambah sungguhpun semburan racun serangga
telah dilakukan. Keadaan ini telah dikaitkan
dengan letusan semula populasi akibat rang-
sangan pembiakan yang tebih tinggi pada
kepekatan submaut di lapangan (Sithananthan,
et al., 1973).
Kajian ini telah dijalankan untuk menen-
tukan kesan racun karbaril terhadap keupayaan
pembiakan A. craccivora Koch.
BAHAN DAN KAEDAH
Kultur Afid
Kultur stok telah dimulakan dengan seekor afid
betina yang dipelihara pada pokok kacang
 YUSOF BIN IBRAHIM DAN WONG KUAN YEOW
tanah, Arachis hypogaea var. Matjan, dalam in-
sektarium pada suhu 24°C - 32°C dan kelem-
bapan bandingan 46% - 90%. Pokok-pokok
kacang tanah ini telah dipastikan bebas dari
serangan Iain-lain serangga dan penyakit. Ke-
segaran kultur stok ini dijaga dengan mengawal
saiz populasi dan menyalinkan pokok-pokok
kacang tanah yang segar setiap empat hari.
Rawatan
Untuk rawatan, nimfa-nimfa instar pertama dari
generasi ketiga yang telah dilahirkan dalam
jangka masa 24 jam telah digunakan.
Sebagai substrat rawatan, pokok kacang
tanah yang subur dan berumur 10 hari telah
digunakan. Setangkai daun yang tidak terasing
dari pokoknya telah dicelupkan ke dalam ban-
cuhan rawatan racun karbaril atau air suling
dan dibiarkan kering udara selama setengah
jam. Bancuhan rawatan racun karbaril (Sevin
85WP) yang digunakan ialah pada kepekatan
0.0106%, 0.0053%, 0.0027% dan 0.0013% serta
air suling yang merupakan rawatan kawalan.
Seterusnya, nimfa instar pertama telah dipin-
dahkan secara cermat dengan berus halus (000
Series Finest Sable, Winsor & Newton) kepada
daun yang telah dirawat. Nimfa afid ini kemu-
diannya membesar di dalam sangkar mikro yang
diperbuat dari dua potongan paip PVC ber-
garis pusat 5 cm setebal 5 mm sebagai rangka.
Sangkar mikro ini ditutupi dengan kepingan
PVC lutsinar nipis yang berlubang-lubang halus
untuk penyaliran udara. Dalam membentuk satu
sangkar mikro, tetali span dikepitkan di antara
dua rangka PVC tersebut untuk mengelakkan
tangkai daun daripada tercedera. Untuk men-
jamin kesihatan afid dan mengelak pertum-
buhan kulat, sangkar mikro ini sentiasa diber-
sihkan dari madu serangga dan wap air.
Data dan Penganalisisan
Setiap hari pada waktu yang sama bilangan afid
yang mati dan yang masih hidup, iaitu data
kemandirian, serta nimfa yang dilahirkan telah
dicatat bagi setiap afid secara individu sehingga
ianya mati. Setelah dibilang, nimfa-nimfa telah
dinyahkan daripada sangkarnya. Prosedur
jadual hayat standard telah diikuti dalam mem-
buat penganalisisan data untuk mengira para-
meter-parameter seperti kadar pembiakan kasar
(GRR), kadar pembiakan bersih (Rf)), kadar
PERTANIKA VOL. 12 NO. 1, 1989
intrinsik pertambahan (r) dan masa purata
generasi (T) (Southwood, 1978). Min daripada
empat replikat yang setiap satunya mengan-
dungi 10 afid bagi setiap rawatan telah di-
bandingkan dengan cara analisis varian dua hala
dan ujian julat berganda Duncan (DMRT).
KEPUTUSAN
Kadar kemandirian terkhusus umur bagi A. rrac-
civora setelah dirawat dengan beberapa ke-
pekatan racun karbaril (Rajah 1) telah didapati
menurun bersamaan dengan kenaikan ke-
pekatan rawatan. Masa tumbesaran nimfa ke
dewasa ialah 4 hari. Bagi rawatan yang terpekat
sekali (0.0106%), hanya terdapat tiga peratus
nimfa afid yang mencapai peringkat dewasa
berbanding dengan 98% dalam kawalan. Se-
terusnya, terdapat lapan peratus mencapai
kedewasaan bagi kepekatan 0.0052%, 40% bagi
kepekatan 0.0027% dan 68% bagi kepekatan
0.0013%. Rawatan pada dua kepekatan tertinggi
juga mengakibatkan umur afid pendek, iaitu 10
dan 12 hari jika dibandingkan dengan 29 hari
dalam kawalan.
Rajah 1. Kemandirian Aphis craccivora Koch yang
dipelihara atas pokok kacang tanah yang
telah dirawat dengan karbaril di dalam
keadaan insektarium.
Secara amnya, corak pembiakan A. crac-
diwaadalah agak sama bagi kesemua kepekatan
rawatan (Rajah 2). Kadar pembiakan didapati
lebih tinggi pada peringkat awal dewasa yang
kemudiannya menjadi semakin berkurangan
mulai umur lapan hari. Puncak pembiakan pada
kesemua rawatan ialah pada hari ketujuh ke-
 KESAN PENDEDAHAN RACUN KARBARIL TERHADAP TUMBESARAN DAN PEMBIAKAN A CRACCIVORA KOCH

cuali afid yang dirawat dengan 0.0106% b.a.

yang mana puncak pembiakannya hanya di-
capai pada hari kesepuluh. Tambahan pula
puncak pembiakannya didapati 41 kali lebih
rendah daripada kawalan. Bagi kepekatan yang
lebih rendah, pendedahan kepada 0.0053%,
0.0027% dan 0.0013% masing-masing menun-
jukkan puncak pembiakan 29, 2.4 dan 1.3 kali
lebih rendah daripada kawalan. Panjangnya
jangka masa pembiakan juga didapati lebih
pendek bagi afid yang dirawat berbanding
dengan kawalan. Pada kepekatan 0.0106% dan
0.0053%, afid hanya berupaya membiak selama
empat dan lapan hari sahaja berbanding de-
ngan 22 hari pembiakan bagi kawalan. Ini ber-
makna masa pembiakannya telah dipendekkan
5.5 dan 3.0 kali berbanding dengan kawalan.
Jadual 1 menunjukkan parameter-para-
meter jadual hayat afid yang telah dirawat dan
dipelihara di atas pokok kacang tanah. Pada
kawalan didapati bahawa kadar pembiakan kasar
(GRR) ialah 47.06 nimfa dan kadar pembiakan
bersih (Ro) ialah 42.13 nimfa. Kedua-dua para- J.2 16 20

meter ini memberikan nilai yang agak tinggi Umur ( H a r i )

dan menunjukkan perbezaan yang bererti se-

tidak-tidaknya bagi afid-afid yang telah dirawat Rajah 2. Kesuburan Aphis craccivora Koch yang
dengan kepekatan karbaril yang melebihi dipelihara atas pokok kacang tanah yang
0.0053%. Begitu juga halnya dengan masa telah dirawat dengan karbaril dalam keadaan
purata generasi (T) yang telah dipendekkan insektarium.
dari 11 kepada 5 hari pada kepekatan 0.0053%
dan 2.6 hari sahaja pada kepekatan 0.0106%
yang mana telah mengakibatkan berkurangnya
kadar pertambahan saiz populasi. Ini amat finit pertambahan (X) populasi yang mencatat-
ketara sekali apabila dilihat dari segi kadar kan nilai kurang dari satu. Parameter masa
intrinsik pertambahan (r) yang mencatatkan ganda dua (DT) tidak menunjukkan perbezaan
pertumbuhan negatif dan juga dari segi kadar yang bererti (P=0.05).
JADUAL 1
Jadual hayat Aphis craccivora Koch yang
dipelihara atas pokok kacang tanah yang telah dirawat dengan karbaril dalam insektarium.

Parameter
Kepekatan
(% b.a.) GRR T r 31 DT

0.0106 2.25a 0.225a 2.61 a - 0.053a 0.25a -17.32 a

0.0053 2.38a 0.450a 5.07a - 0.003 ab 0.46a 3.02a
0.0027 41.46b 15.401 b 11.06b 0.250bc 1.28b 2.81 a
0.0013 43.93 b 27.670c . 11.29b 0.290c 1.34b 2.42a
Kawalan 47.06b 42.130d 11.31 b 0.330c 1.39b 2.12a

Min-min dalam ruang yang sama dan diikuti dengan huruf yang sama adalah tidak mempunyai perbezaan yang bererti
pada paras P < 0.05 setelah diuji dengan ANOVA dua hala dan DMRT.

PERTANIKAVOL. 12 NO. 1, 1989 45

 YUSOF BIN IBRAHIM DAN WONG KUAN YEOW
PERBINCANGAN
Keputusan yang diperolehi telah menunjukkan
bahawa dalam keadaan ambien pada suhu 28
± 4°C, kelembapan bandingan 68 ± 22% dan
tiada gangguan dari faktor-faktor lain yang
mengancam perkembangan populasi, maka A.
craccivora Koch berupaya melipat gandakan saiz
populasinya dalam masa dua hari sahaja. Dalam
masa yang sesingkat ini nimfa-nimfa yang di-
lahirkan berkembang cepat dan hanya me-
merlukan 4-5 hari untuk menjadi dewasa. Tam-
bahan pula, perkembangan ovari dan embrio
bermula semasa peringkat nimfa lagi (Elliott
dan McDonald, 1976), maka kelahiran progeni
dapat bermula pada hari pertama afid itu men-
jadi dewasa (Rajah 2). Jika dilihat kepada nilai
Ro = 42.13 dan T = 11.31 (Jadual 1), maka
secara teori, berpunca dari seekor afid sahaja
tidaklah mustahil saiz populasi boleh bertam-
bah me-lebihi 132 juta ekor dalam masa 56.5
hari. Sebaliknya, populasi afid yang telah did-
edah-kan kepada 0.005% karbaril telah men-
jadi semakin berkurangan pada setiap generasi
yang mana akhirnya akan menjadi pupus. Oleh
itu, keputusan yang diperolehi dalam kajian ini
telah menunjukkan bahawa karbaril mampu
ber-tindak sebagai perencat perkembangan
populasi afid yang baik, iaitu sependapat dengan
Sarup et al (1969).
Semasa melakukan semburan karbaril di
lapangan, biasanya akan terdapat afid yang
terlepas dan tidak bersentuhan dengan sem-
buran pada dos maut, maka ia akan terus hidup.
Sungguhpun begitu, ini sepatutnya tidak perlu
dibimbangkan kerana pada kepekatan yang
rendah, iaitu mungkin dari sisa semburan yang
lepas, sudah mencukupi untuk membunuh atau
merencatkan tumbesaran nimfa-nimfa muda.
Adalah menarik apabila dilihat bahawa
pendedahan kepada kepekatan 0.0106% telah
mengakibatkan masa prapembiakan diper-
lanjutkan, iaitu tiga hari lebih lewat daripada
kawalan. Secara tidak langsung keadaan ini akan
melewatkan kerosakan pada tanaman dan
melambatkan pertumbuhan populasi afid.
Dengan itu keputusan kajian ini juga menun-
jukkan bahawa racun karbaril pada kepekatan
0.01% adalah memadai untuk pengawalan
serangga perosak A, craccivora.
46 PERTANIKAVOL. 12 NO. I, 1989
PENGHARGAAN
Terima kasih diucapkan kepada Universiti
Pertanian Malaysia di atas kebenaran
menggunakan kemudahan-kemudahan yang
terdapat di insektarium Jabatan Perlindungan
Tumbuhan.
YUSOF BIN IBRAHIM
WONG KUAN YEOW
Jabatan Perlindungan Tumbuhan
Fakulti Pertanian
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan,
Malaysia.
RUJUKAN
ELLIOTT, HJ. dan FJ.D. MCDONALDS. 1976. Re-
production in a Parthenogenetic Aphid, Aphis
craccivora Koch. Embryology, Ovarian Develop-
ment and Fecundity of Apterae and Alate. Auat.
J. ZooL 24: 49-63.
HALL, F.R. 1979. Effects of Synthetic Pyrcthmids
on Major Insect and Mite Pests of Apples. /. Econ.
EntomoL 72: 441-446.
JONES, V.P. dan M.P. PARRFLLA. 1984. The Subli•-
thai Effects of Selected Insecticides on Life Table
Parameters of Panonyckus citri (Acari: Tetranychi-
dae). Can. EntomoL 116: 1033.
LUCKEY, T.D. 1968. Insecticide HormoHgosis. /.
Earn. EntomoL 6 1 : 7-12.
RADKE, S.G., W.G. YKNDOL, dan A.W.
BENTON. 1972. Studies on Parthenogenetic
Viviparous and Sexual Forms of the Cowpea aphid,
Aphis craccivora Koch. Indian J. EntomoL 34: 319-
324.
SARUP, P., D.S. SINGH, V.S. SRIVASTAVA, P. SIRCAR.
dan RATTAN LAL. 1969. Laboratory Evaluation
of Different Pesticides as Contact Poison Against
the Adult of Aphis craccivora Koch Infesting Pea
Crop. Indian]. EntomoL 31: 311-320.
SITHANANTHAM, S., S. JAYARAJ, dan T.R. SUBRA
MANlAM. 1973. Some Changes in the Biochemi-
cal Status of Cotton Plants due to Systemic Insec
ticidal Protection, in Relation to Resurgence oi
the Aphis gossypii Glow (Homop.: Aphididae)
Madras Agric. J 60: 512-518.
SOUTHWOOD, T.R.E. 1978. Ecological Method;
with Particular Reference to the Study of Insec
Population. New York: Halsted Press. 524 pp.
(Received 2 July, 1988
Pertanika 12(1), 47-51 (1989)

COMMUNICATION II
The Rate of Litter Production in Mangrove Forest at Siar Beach,
Lundu, Sarawak
ABSTRAK
Penyelidikan kadar pengeluaran sesampah huian bakau telah dilakukan di Pantai Siar, Lundu, Sarawak.
Huian ini mempunyai sembilan spesies pohon; spesies pohon yang dominan ialah Rhizophora mucronata,
sementara R, apiculata sebagai spesies kodominan. Kadar pengeluaran sesampah ialah 5. 72 t/ha/tahun,
dan dari jumlah ini 4.49 t/ha/tahun (78.5%) adalah serasah daun. Corak kadar pengeluaran sesampah
dan serasah daun berubah antara satu bulan dengan lainnya. Pengeluaran sesampah antara bulan April
hingga Jun dan Disember hingga Februari lebih banyak berbanding dengan bulan-bulan lain.

ABSTRACT
A study of litter production was conducted at Siar Beach, mangrove forest, near Lundu, Sarawak. There are
nine species of trees in this forest of which Rhizophora mucronata is the dominant and R. apiculata is the
co-dominant species. The rate of litter production xoas 5.72 t/ha/year, and of this 4.49 t/ha/year (78.5%)
was leaf litter. 'The production of litter fluctuates being higher during April to June and December to February
compared to other months.

INTRODUCTION Gong et al. (1984) in Matang mangrove forest,

Mangrove trees serve many functions. The
ecological functions of mangrove as a land
builder and coastline stabilizer have been
described by McNae (1974) and de la Cruz
(1979). Heald and Odum (1970) described
mangrove as a spawning and nursery area for
fishes and prawn. These functions, are closely
related to the primary productivity of the
mangrove. Several authors have conducted
studies on the relationship of primary produc-
tivity of mangroves and its ecological and eco-
nomic functions (e.g. Golley et al. 1962; Heald
and Odum 1970; Odum 1971). They found
that most of the organic debris that enriched
the estuary were derived from adjacent mang-
rove forest. Thus, litter production plays an im-
portant role in maintaining the fertility of the
mangrove ecosystem and supply of food mate-
rial to the faunal life (Srivastava 1980). How-
ever, the rate of litter production varies from
place to place, and between species dominating
the site (Mathias 1974). Such differences can
be seen from the work of Heald (1971) in
Florida, Christensen (1978) in Thailand and
West Malaysia.
The production of litter of the mangrove
forest at Siar Beach, near Lundu was studied in
order to understand more about the producti-
vity of mangrove ecosystems in Sarawak.
Study Area
The study area is located at Siar River Estuary,
between latitudes 109° 52' E to 109° 53' E and
longitudes 1° 44' to 1° 45' N (Wolfenden and
Haile 1963). The estuary is about 12 km N - E
of Lundu (Fig. 1). The river is about 2 m wide
at low tide and 4 m wide at high tide.
The area is dominated by Rhizophora
mucronata and R. apiculata is co-dominant
(Saberi 1984). The other species that can be
found in the area are Bruguiera gymnorrhiza, B.
parviflora, B. cylindrica, Avicennia alba, Lumnitzera
sp., Nypa fruticans, Oncosperma sp. and a fern
Acrosticum sp.
The soils of the area derived from biotite
adamellite rocks. The climate is reasonably wet,
highest precipitation occurring in January, and
lowest in March. The total precipitation is 4260
mm per year. The temperature varies between
 SABERI OTHMAN

SOUTH CHINA SEA

Kuching 1 ° 4 1 ' N-

Fig. 1. Star beach of Lundu, Sarawak.

48 PERTANIKA VOL. 12 NO. I, 1989

 THE RATE OF LITTER PRODUCTION IN MANGROVE FOREST AT SIAR BEACH, LUNDU, SARAWAK
25° - 28°C, the lowest being in January and
higher in May (Fig. 2).
Fig. 2. The climatic diagram for Kuching, Sarawak.
MATERIALS AND METHODS
Plant litter was collected in a trap made of sail-
cloth, with a square wooden-frame opening
measuring 50 x 50 cm. The trap was placed lm
above the ground. Fifteen such litter traps were
sited randomly beneath the trees, and their
contents were emptied at the beginning of each
month. The litter was sorted into leaves and a
miscellaneous component. The litter was washed
and oven-dried at 105°C to constant weight.
RESULTS AND DISCUSSION
The suitable size and form of litter-trap for
sampling litter-fall in a tropical rain forest is not
known. A study of previous work shows that a
varied size of litter-traps were used. The small-
est was 45 x 45 cm (Woodroffe 1982) and the
largest was 200 x 200 cm (Duke et al. 1981).
The form of the trap also shows variation. Mason
(1970) had used round shaped traps instead of
square, and Kira et al. (1967) used a long belt
of plastic net as the trap. In this study a trap
with a square opening measuring 50 x 50 cm
was used. This size is within the range of the
trap used by previous workers, and easy to
handle while in the forest.
PERTANIKA VOL. 12 NO. 1, 1989
The total small litter and leaf litter pro-
duction varies between forest and climatic
regions. Generally leaf litter accounts for more
than 50% of the total litter produced. However,
the amounts of this litter may vary depending
on the size of the traps used. For some forests,
particularly mangrove, leaf litter could reach as
high as 80%. A similar high percentage (78.5%)
of leaf litter was detected in the mangrove forest
at Siar Beach (Fig. 3).
50"
40
1 | Leaf-litter
p^l Miscellaneous
Fig. 3. The pattern of litter fall for 1983 and 1984 in a stand
of Mangrove forest, Siar Beach, Lundu, Sarawak.
The mean values obtained for the mixed
mangrove forest at Siar Beach are 5.74 and 4.49
t/ha/year for total small litter and leaf litter
respectively. These values are significantly lower
than those obtained for many tropical rain forest
(e.g. Nye 1961; Bullock 1973; Gong et al. 1984).
The leaf litter of this forest is about 40% lower
compared to the values obtained for other
mangrove forests. This could be due to the fact
that the mangrove forest at Siar Beach is not as
dense compared to other forests. Such situations
have been highlighted by Gong et al (1984)
for Matang mangrove forest, and Woodroffe
(1982) for Tuff Crater, mangrove forest, Auck-
land New Zealand. The density of trees at Siar
Beach is 166/ha (Saberi, 1984), and this value
49
 SABERI OTHMAN

is about four times lower compared to other fo- REFERENCES

rests such as at Banyu Asin, Sumatera (Soekardjo BULLOCK, J.A. 1973. Cited in Golley,
and Kartawinata 1979), Sabah, Malaysia (Liew R.B.I978.Gross and Net Primary Production and
1977) and in Florida, U.S.A. (Pool et al. 1977). Growth Parameters, pp. 233-248, in Tropical Forest
However, there is no data available for compa- Ecosystems. UNESCO Natural Resources Research
rison on the relationship between density and Vol. XIV, UNESCO-NEP, France.
litter production in mangrove forests. CHRISTENSEN, B. 1978. Biomass and Primary Pro-
Litter production increased in March and duction of Rkizopkora apiculata Bl. in a Mangrove
April and reached its highest rate in May. It in Southern Thailand. Aquatic Botany 4: 43-52.
subsequently dropped in June, July and August CHRISTENSEN, B. and WIUM - ANDERSEN. 1971.
and remained constant till November. It then Seasonal Growth of Mangrove Trees in Southern
increased again in December and reached its Thailand. I. The phenology of Rhizophora apicu-
second peak in February. A similar pattern in lata Bl. Aquatic Botany 3: 281-286.
litter fall has been reported for Pasoh (Lim CRUZ, A.A. de la. 1979. The Functions of Mang-
1978), Matang mangrove forest (Gong et al. roves. Symposium on Mangrove and Estuarine Vege-
1984) and R. apiculata in Phuket, Thailand tation in Southeast Asia. Universiti Pertanian
(Christensen and Wium-Anderson, 1977). The Malaysia, Serdang, Selangor, 25th - 28th April,
1978.
high production in May coincide with high
temperature and draught. Gill and Tomlinson DUKE, N.C., J.S. BUNT, and W.T. WILLIAM. 1981.

(1969 & 1971) in their studies of the red mang- Mangrove litter-fall in NE Australia I. Annuals
Totals by Component in Selected Species. Austra-
rove in South Florida, report peak rates of leaf
lian Journal of Botany 29: 547-553.
fall and growth during summer months when
air temperature and incident light were at their GILL, A.M. and P.B. TOMLINSON. 1969. Studies on
the Growth of Red Mangroves {Rhizophora mangle
annual peaks. Similar results have been reported
L.). I. Habitat and General Morphology. Biotro-
by Heald (1971). Snedaker and Lugo (1973)
pica 1(1): 1-9.
found indications that leaf fall increases during
GILL, A.M. and P.B. TOMLINSON. 1971. Studies on
dry periods. According to Lugo and Snedaker
the Growth of the Red Mangrove (Rhizophora
(1975) the timing of these events has a signifi-
mangleL.). II. Growth and Differentiation of Aerial
cant interaction with the hydrologic budget and Roots. Biotropica 3: 109-24.
primary productivity of the forest community.
GOLLEY, F.B., M.T. ODUM, and R.F.
The higher production in February cannot be
WILSON. 1962. The Structure and Metabolism
explained in terms of biological processes.
of a Puerto Rican Red Mangrove Forest in May.
However, its does coincide with heavy winds of Ecology. 43: 9-13.
the South China Monsson and rainfall. Occa-
GONG, W.K., J.E. ONG, C.H. WONG and G.
sional storms have been reported to coincide
DHANARAJAN. 1984. Productivity of Mangrove
with increased litter fall (e.g. Pool et al. 1975; Trees and its Significance in a Managed Man-
Goulter and Allaway 1979). grove Ecosystem in Malaysia. Asian symposium on
Production of litter is relatively continu- Mangrove Environment: Research and Management
ous, thus supplying the detritus food chain. The (Eds, Soepadmo, E, Rao, A.N. and Macintos, DJ.)
fraction exported to adjacent sea areas is not pp. 216-225.
known, but many leaves are carried away by GOULTER, P.F.E. and W.G. ALLAWAY. 1979. Litter
tides, especially at spring tides, before any Fall and Decomposition in a Mangrove Stand,
degradation takes place. Aricennia marina (Forsk.) Vierh., in Middle Har-
bour, Sydney. AusLj. Mar. Freshwater Res. 30: 541-
SABERI OTHMAN 546.
HEALD, EJ. 1971. The Production of Organic
DepartTnent of Biology,
Detritus in a South Florida Estuary. Univ. Miami
Faculty of Science and Environmental Studies
Sea Grant. Tech. Bull. No. 6, 100 pp.
Universiti Pertanian Malaysia
HEALD. EJ. and W.E. ODUM. 1970. The Contribu-
43400 UPM Serdang Selangor Darul Ehsan,
tion of Mangrove Swamps to Florida Fisheries.
Malaysia,
Proc. Guld and Carrib. Fish. Inst. 22: 130-135.

PERTANIKA VOL. 12 NO. 1, 1989

 THE RATE OF LITTER PRODUCTION IN MANGROVE FOREST AT SIAR BEACH, LUNDU, SARAWAK
KIRA, T., H. OGAWA, K. YODA, and K.
OGINO. 1967. Comparative Ecological Studies
on Three Main Types of Forest Vegetation in
Thailand. IV. Dry matter Production with Special
Reference to the Khao Chong Rain Forest. In T.
Kira and K. Iwata (Editors), Nature and Life in
Southeast Asia, Vol. 5, Flora and Fauna Research
Society, Kyoto, Japan, pp. 149-177.
LlEW, T.C. 1977. Mangrove Forests of Sabah. In
Proc. workshop on Mangrove and Estuarine Vegetation
Eds. Srivastava and Razali Kader pp. 6-31. Kuala
Lumpur.
LiM, M.T. 1978. litter-fall and Mineral Nutrient
Content of Litter in Pasoh Forest Reserve. Malay.
Nat J, 30: 375-80.
LUGO, A.E. and S.C. SNEDAKER. 1975. Properties
of a Mangrove Forest in Southern Florida. Proceed-
ings of International Symposium on Biology and Man-
agement of Mangroves. 1: 170-212.
MASON, C.F. 1970. The Structural and Temporal
Patterns of Litter Fall in Wytham Woods, near
Oxford, England. Oecologia 5: 215-39.
MATHIAS,J.A. 1974. The Possible of Leaf Fall and
Decomposition in Malaysian Mangrove Swamps.
I.B.P. Synthesis meeting, Kuala Lumpur.
MCNAE, W. 1974. Mangrove Forests and Fisher-
ies. FAO Publ. IOFC/DEV/74/34., FAO, Rome,
p. 7.
NVE, P.H. 1961. Organic Matter and Nutrient 333-
346.
ODUM, W.E. 1971. Pathways of Energy Flow in a
South Florida Estuary. Univ. Miami, Sea Grant
Tech. Bull. No. 7.
POOL, D.J., A.E. LUGO and S.C.
SNEDAKER. 1975. Litter Production in Mang-
rove Forest of Southern Florida. In Proceeding of
PERTANIKA VOL. 12 NO. 1, 1989
International Symposium on Biology and Management
of Mangroves, eds. G. Walsh, S. Snedekar and H.
Teas. 1: 213-237.
POOL, D.J., S.C. SNEDAKER, and A.E.
LUGO. 1977. Structure of Mangrove Forests in
Florida, Puerto Rico, Mexico and Costa Rica.
Biotropica 9(3): 195-212.
SABERI, O. 1984. Structure of Mangrove Vegeta-
tion at Siar Beach, Lundu, Sarawak. Pertanika7(2):
91-99
SOEKARDJO, S. and K. KARTAWINATA. 1979. Mang-
rove Forests of Banyu-Asin, Musi River Estuary,
South Sumatera, Indonesia. In Mangrove and Es-
tuarine Vegetation in Southeast Asia. Sym. Edited by
P.B.L. Srivastava, Abdul Manaf Ahmad, G. Dha-
narajan and Ismail Hamzah. Biotrop Pub. No. 10.
Bogor.
SNEDAKER, S. and A. LUGO. 1973. The role of
Mangrove Ecosystems in the Maintenance of
Environmental Quality and a High Productivity
of Desirable Fisheries. Atlanta, Ga.: U.S. Bureau
of Sports Fisheries 8c Wildlife. NTIS Springfield.
SRIVASTAVA, P.B.L. 1980. Research Proposals for
Mangrove Vegetation in Malaysia. Workshop on
Mangrove and Estuarine Vegetation, (Ed. P.B.L.
Srivastava and Razali Abdul Kadir). pp. 64-75.
Kuala Lumpur.
WOLFENDEN, E.B. and N.S. HAILE. 1963. Sematan
and Lundu Area, West Sarawak. Report I. Geo-
logical Survey Department, Sarawak, pp.9-11.
WOODROFFE, C D . 1982. Litter Production and
Decomposition in the New Zealand Mangrove,
Avicennia marina var. resinifera. N.Z.J. Mar. Fresh-
water Res. 16(2): 179-188.
(Received II March, 1987)
51
Pertanika 12(1), 53-58 (1989)

COMMUNICATION III
Diseases in Dogs: Necropsy Observation
ABSTRAK
Kematian anjing disebabkan oleh penyakit atau keadaan tak normal terutamanya dalam sistem pansistemik,
diikuti oleh sistem penghadaman dan sistem pemafasan. Dalam kategori pansistemik, kebanyakan kematian
disebabkan oleh keracunan dan penyakit distemper. Penyakit utama pada sistem penghadaman adalah
ansylostomiasis manakala penyakit utama pada sistem pemafasan adalah bronkopneumonia.

ABSTRACT
Morbidity in dogs was associated mainly with diseases or abnormalities in the pansystemicy digestive and
respiratory systems in that order. In the pansystemic category, poisoning and canine distemper were most
common. Ancylostomiasis and bronchopneumonia were the main findings in the digestive and respiratory
systems respectively.

INTRODUCTION ii) the reason for euthanasia. These were then

Dogs form a significant proportion of the total classified according to body systems. Cases whic h
number of cases submitted for necropsy to the were not finalised and cases with no diagnosis
Faculty of Veterinary Medicine and Animal were not included.
Science, Universiti Pertanian Malaysia (UPM).
From June 1979 to August 1984, dogs made up Results and Discussion
15.9% of the total number fo carcasses submit- Out of 2376 cases submitted for necropsy over
ted for necropsy (Chooi 1985), Apart from the study period, dogs constituted 331 cases
specific case reports, there has been no docu- (13.9%). Final diagnosis was made on 219 of
mentation on diseases of dogs in Malaysia. This these 331 cases. The rest were incomplete cases.
paper summarizes the post mortem findings Pansystemic (18.3%), digestive (16.4%) and
in dogs submitted for necropsy to UPM. respiratory (13.2%) systems were most com-
monly the sites of disease (Table 1).
Materials and Methods TABLE 1
Necropsy records in the Faculty of Veterinary Involvement of body systems in dogs examined
Medicine and Animal Science, UPM over the post mortem at UPM
5-year period from January 1982 to December
System No. of cases <%)
1986 were analysed for disease conditions in
dogs. Cases submitted were either from the small Pansystemic 40 (18.3)
animal clinic in UPM or from private practitio- Digestive 36 (16.4)
ners in Kuala Lumpur and Petaling Java. Most Respiratory 29 (13.2)
of the cases were dead animals or animals that Musculoskeletal 21 (9.6)
were euthanised because of poor prognosis. Cardiovascular 21 (9.6)
The majority were pet animals; a few were Renal 21 (9.6)
working dogs. Nervous 12 (5.5)
Hepatic " (5.0)
Carcases were subjected to a routine ne- 10
Reproductive (4.6)
cropsy. Ancillary studies in histopathology, Integumentary 6 (2.7)
bacteriology and virology were conducted when Miscellaneous 12 (5.5)
necessary. Analysis of the cases was based on i)
disease, or disease condition causing death and Total 219 (100.0)

•
 TABLE 2
Disease conditions in the pan systemic, digestive and respiratory systems
Pansystemic Digestive Respiratory

Condition No of Condition No. of Condition No. of

cases (%) cases (%) cases (%)

Paraquat poisoning 10 (25.0) Ancylostomiasis 15 (41.6) Bronchopneumonia 14 (48.0)

Canine distemper 8 (20.0) Parvovirus Intersitial pneumonia 2 (7.0)
Starvation 4 (10.0) enteritis (30.5) Fibrinous pneumonia 2 (7.0) Q
Organophosphate
poisoning 4 (10.0)
Hemorrhagic
enteritis 3 (8.3)
Aspiration pneumonia
Acute pneumonia
1
1
(3.5)
(3.5)
I
H
Lymphosarcoma 3 (7.5) Granulomatous Hemorrhagic pneumonia 1 (3.5)
Heat stroke 2 (5.0) esophagitis 2 (5.6) Mucopurulent rhinitis 2 (7.0)
Warfarin poisoning 1 (2.5) Acute pancreatitis 2 (5.6) Diaphragmatic hernia 1 (3.5) 2

I Strychnine poisoning
Lead poisoning
Septicemia
Thrombocytopenic
1
1
1
(2.5)
(2.5)
(2.5)
Uremic gastritis
Ulcerative gastritis
Gastric torsion
9
1
1
1
(2.8)
(2.8)
(2.8)
Nasal carcinoma
Asphyxiation
Bronchogenic carcinoma
1
3
1
(3.5)
(10.0)
(3.5) >
purpura 1 (2.5) w
Ehrlichiosis 2 (5.0) >
Toxoplasmosis 1 (2.5)
Poisoning
(undetermined) 1 (2.5) s
6
Total 40 (100.0) 36 (100.0) 29 (100.0)

•
 TABLE 3
Disease conditions in the musculoskeletal, renal and cardiovascular systems

Musculoskeletal Renal Cardiovascular

Condition No. of Condition N o . of Condition No. of

cases < % )
cases (%) cases (%)
Fracture (axial Nephritis 13 (62.0) Dirofilariasis 18 (85.0)
& appendicular) 9 (42.8) Urolithiasis 2 (9.0) Myocardial degeneration o
Hip dysplasia 6 (28.4) Leptospirosis 2 (9.0) 8c necrosis 1 (5.0)
Rickets 1 (4.8) Oxalate nephrosis 1 (5-0) Congestive heart 2 (10.0)
Arthritis 1 (4.8) Renal hemorrhage , 1 (5.0) failure
PERT

Myositis 1 (4.8) Membranuos 2;

Spinal column glomerulopathy •t*n~~*' ... T — (5.0) - — - — 0
g luxation 1 (4.8) Pyelonephritis 1 (5.-0) ;•'• • . ;^U i ;
- 0
Fibrosis of joint 1 (4.8)
<: Osteosarcoma 1 (4.8)
2:
w
o 0
50
12 NO 1 . 1 9

OPSY
Total 21 (100.0) '..fifi • i ' f ^
21 (100.0) 21 (100.0)

•har ^ : : : •-
u
oo 1
to *;--'Ail-->ty •>;/•' : •
:• I:'.- , n : - • • • . • . . . . • - , ' . - •••• . , ;
:, • • ' .. . \ i y ,• .•

| i*.ff< f f r ; v*}'--•- ::
•• hfu* :.' *T' f,. • '>.~'*i • < : '• '.I
1 : (
' :. ' . j . . . .

; .-•';>'•••; i • • '
•' - f. '•-; ; • 3 : • • ' • • * ; - '•«**• - . . '-' - y

* : j
t
. A " J - • • • : ; • ;•'•
'- .*.«] ifdiir.fr/ • _' c r x . V'S* *•'•>'/:
•;e f .:!^.-ir:-v.v*:."-»'.K ' --; :
.' -- "•
• -L'

?: . . .; ^
K . - \ ~ < y . ' : • • . . • * •

i-o.'xjwsw . . • • ' • • ' ;

- i
•• : i_-;>rjKf;.£io<( • v .; ' . ••

~; i
 TABLE 4
Disease conditions in the nervous, hepatic and reproductive systems.
Nervous Hepatic Reproductive
No. of
Condition No. of Condition No. of Condition cases
cases (%) cases (%)
Meningitis 4 (33.4) Storage disease 3 (27.0) Transmissible
Spinal dural Necrosis • 2 (18.5) venereal tumour 8 (80.0) r>
ossification 3 (25.0) Trauma 1 (9.0) Ovarian cyst
Degenerative Jaundice 1 (9.0) adenocarcinoma 1 (10.0)
myelopathy 2 (16.7) Congestion 1 (9.0) Stillbirth 1 (10.0)
Spinal cord . Hepatitis 2 (18.5)
hemorrhage ^ ^ ^ 1 ^ (8 3) ._, Hepatocellular .^^^
Cerebral focal carcinoma 1 (9.0)
necrosis 1 (8.3)
Intervertebral disc
prolapse i (8.3) J
§ a
Total 12 (100.0) 11 (100.0) 10 (100.0) >

I f :
- • : ; • : • • \ •:• ] .
X
tn
;
' ; ' " ' •'; . '; '• • " . . ! • • . C < - ' ^ ^ [ '

• • ' • • " - ' * • • • • - ' - • •••••• • - . - • . - • , - ^ . ^ ^ , , , . j - 3 . . . . . . . . ( . . _ £ & ' - . , . . . : , , . . . . -

• • • • • ' ••, - / • < ~ - * f < # . * " < ; , • • £ /

i \.-r-./t.v. '. , i . > i : t j =• i . v - - * . i ' > * .••• ';;t ;•-;*'^ ^ n ^ ^ ^ C J C : : ^ :.: - - ' ' I >•..?: ,:;M^W;.:,;*? > T( rri. *A>('-.*. •
 DISEASES IN DOGS: NECROPSY OBSERVATION

Of the 40 cases classified under pansys-
temic, 18(45%) were cases of poisoning (Table
2). Ten of these were due to paraquat; six of
which have been described elsewhere (Chooi 8c
Ibrahim 1985; Chooi et al 1986). Canine dis-
temper accounted for 20% (8) of the pansys-
temic diseases. The low prevalence is probably
because most pet dogs and all working dogs are
vaccinated against canine distemper which is
enzootic in this area. Death occurs mainly in
non-vaccinated or improperly vaccinated ani-
mals.
In the digestive system, ancylostomiasis
was most common followed by parvovirus
enteritis (Table 2). Cases of parvovirus infec-
tion have been reported elsewhere (Noor et al.
1980;Omar^a/. 1980; Sheikh-Omar et al 1985).
Three cases of hemorrhagic enteritis not due
to parvovirus infection were diagnosed. One of
these was due to clostridial infection (Chooi 8c
Netto 1986) while the cause of the other two
cases was undetermined. Granulomatous eso-
phagitis due to Spirocerca lupi infestation was
found in two dogs.
Bronchopneumonia accounted for
48% of the cases in the respiratory system.
Bacteria isolated included Staphyloccocus and
Klebsiella species. There was a likelihood that
these cases were secondary to viral infections
although histological examination of the lung
in most cases failed to reveal inclusion bo-
dies.
Fractures of the axial and appendicular
skeleton were caused by trauma - usually in
road accidents. These animals, and those with
hip dysplasia, were usually euthanised because
of poor prognosis. Dirofilariasis was the major
problem in the cardiovascular system whilst
subacute to chronic interstitial nephritis was
the main reason for euthanasia or death in the
renal system (Table 3).
In the reproductive system, transmissible
venereal tumour (TVT) was the major reason
for euthanasia (Table 4). Chooi (1985) reported
that TVT was the most common tumour type
in dogs and 91.7% of cases in that report in-
volved the penis or vagina. In addition, tumours
of the mesenchyme, skin and adnexa were of
low prevalence in the present study because
these cases were usually submitted in the form
of biopsies, and were not included here.
There were four cases of meningitis in
the nervous system. Two of these were suppu-
rative. The first case was an 8-week old, female
spitz. The dog was depressed and had nystagmus
prior to death. Staphylococcus aureus was re-
covered from the meninges. The second case
of suppurative meningitis was in a 1-year old
male German Shepherd crossed dog. There was
obvious inflammation of the meninges with ad-
hesions to the cranium at gross examination.
The other two cases were non-suppurative but
further deductions were not possible due to
auto lysis of the brain :

TABLE 5
Disease conditions in the integumentary system and miscellaneous conditions

Integumentary Miscellaneous

Condition No. of cases Condition No. of cases

Mast cell tumour 1 Anaesthetic death 3
Chronic demodecosis 1 Sudden death 2
Histiocytoma 1 Abdominal rupture (post
Dermatitis 1 laparotomy) ]
Squamous cell Lymphadenopathy
carcinoma 1 Anaemia ]
Pediculosis 1 Otitis externa ]
Acute hemorrhage
(post-spay) •

Bilateral cataract :

Hyperparathyroidism ]

Total 6 12

PERTANIKA VOL. 12 NO. 1, 1989 57

 K.F. CHOOI, T. PANDIYARAJA AND A.R. SHEIKH-OMAR

ACKNOWLEDGEMENTS CHOOI, K.F., G.K. DHALIWAL and S. SALMIYAH. 1986.

We thank the pathologists whose cases formed More Cases of Paraquat Poisoning in Dogs. Kajian
a part of this report and Ms. Low Lai Kim for Veterinar 18(1): 94.
typing the manuscript. CHOOI, KJF. and A. NETTO. 1986. A Case of Canine
Hemorrhagic Gastroenteritis. Kajian Veterinar
K.F. CHOOI, 18(1): 89-90.
T. PANDIYARAJA NOOR F., A.L. IBRAHIM, S.E. TAN and CM. LAI. 1980.
A.R. SHEIKH-OMAR A Case of Parvovirus Infection in a Dog. Kajian
Veterinar 12(2): 39-41.
OMAR, A.R, J.Y.S. LEE, and L.T. NO. 1980. Hae-
Faculty of Veterinary Medicine and Animal morrhagic Enteritis of Probable Parvovirus Ori-
Science, gin in Dogs. Kajian Veterinar 12(2):63-66.
Universiti Pertanian Malaysia, SHEIKH-OMAR, A.R., B.Y. CHENG, A.L. IBRAHIM and
43400 UPM Serdang, Selangor Darul Ehsan, K.F. CHOOI. 1985. Parvovirus Enteritis in Two
Malaysia. Dogs. Malays. Appl BioL 14(l):31-33.

REFERENCES
CHOOI, R.F. 1985. Review of Neoplastic Cases of
Domestic Mammals Diagnosed at UPM (1979-
1984). Kajian Veierinar 17(1): 35-42.
CHOOI, K.F. and R. IBRAHIM. 1985. Suspected
Paraquat Poisoning in Four Dogs. Kajian Veterinar
17(2): 147-150. (Received 17 December, 1987)

PERTANIKAVOL. 12 NO. 1, 1989

 Section II
Physical Sciences
Pertanika 12(1), 59-63 (1989)

Chemical Characterisation of Podzols in Sabah, East Malaysia

MARCUS JOPONY and TAN YING CHIU
Faculty of Science and Natural Resources
Universiti Kebangsaan Malaysia, Sabah Campus
88996 Kota Kinabalu, Sabah, Malaysia.

Key words: Podzol; horizon; horizon Bh.

ABSTRAK
Analisis kimia telah dilakukan terhadap tiga sampel podzol dari daerah Tambunan dan Keningau, Sabah.
pH, kandungan karbon organik, muatan pertukaran kation, amaun kation bolehtukar serta amaun logam
yang terekstrak dengan HCI 1.0 M adalah berbeza antara horizon. Perbezaan tersebut adalah paling hetara
antara horizon E dengan horizon Bh. Kecuali pH, nilai bagi setiap parameter di atas adalah terendah dan
tertinggi masing-masing pada horizon E dan Bh.

ABSTRACT
Chemical analyses were carried out on three podzol samples from the Tambunan and Keningau districts in
Sabah. The values ofpH, organic carbon content, cation exchange capacity and HCI extractable metals vary
distinctly between horizons. Such variation was most significant between horizons E and Bh and except for
pH, the values for the above parameters were lowest and highest in horizons E and Bh respectively.

INTRODUCTION The chemical characteristics of podzols

Podzols can be found mainly in humid temper-
ate and humid tropical areas, particularly in
areas where the parent materials are predomi-
nantly of quartz or silicious minerals. In Malay-
sia, podzols are found in Peninsular Malaysia,
Sarawak and Sabah. In Sabah, significant areas
occur in the Tambunan, Keningau and Sipi-
tang districts (Acres et al 1975) The general
characteristics of podzols have been well des-
cribed (McKeague et al 1985; Fitzpat-
rick 1983; Burnham 1984; Duchau-
four 1982; Russell 1973; Muir 1961). The
profile of a podzol is generally characterised by
its clearly defined horizons. An intensively wea-
thered bleached (eluvial) horizon, E, is located
immediately below the upper horizon A. Below
E is a dark coloured illuvial horizon, Bh, also
known as the spodic horizon. A reddish brown
coloured hori-zon, Bh/s, is sometimes found in
between Bh and the parent material (horizon
C), Podzols are basically differentiated from
other soils by the presence of a spodic horizon.
Often the spodic (Bh) horizon is overlain by an
albic (E) horizon.
are widely reported in several works (McKeague
et al. 1985; Torrent and Gomez-Martin 1985;
Evans 1982; Forth 1978; Townsend 1977;
Mohr et al. 1972; Andriesse 1969;
Burnham 1968; Klinge 1965). However, pub-
lished chemical data of Malaysian pod-zols is
very limited and confined mainly to podzols in
West Malaysia.
Therefore the following study was carried
out to assess the chemical characteristics of
podzols in Sabah, East Malaysia. The resulting
data would be a valuable addition to the exist-
ing data on podzols in Malaysia and elsewhere.
MATERIALS AND METHODS
The podzol samples were obtained from Ken-
ingau (PI and P2) and Tambunan (P3) dis-
tricts in Sabah, Malaysia. The soils are of the
Kapayan Association and Brantian Association
respectively, both having terrace landform and
alluvium parent material (Acres et al. 1975).
Road construction works currently going on
along the Kota Kinabalu - Tambunan - Kening-
au highway exposed some areas where the soil
 MARCUS JOPONY AND TAN YING CHIU

was found to be podzols. The podzols at all
sampling sites have distinct A, E, Bh, Bh/s and
C horizons of varying thickness. Soil samples
were taken from the respective horizons with-
out much difficulty since the sites are located
along the highway.
All soil samples were air dried and only
the < 2mm fractions were used in the chemical
analyses. All analyses were carried out in dupli-
cates. Soil pH was measured in a 1:2 suspension
in distilled water using a pH meter. Organic
carbon was determined according to the Walkley
- Black method as described by Nelson and
Sommers (1982). Cation exchange capacity
was determined by 1.0 M ammonium acetate,
and the exchangeable basic cations in the
leachate was measured by atomic absorption
spectrophotometry. Exchangeable Al and H
were extracted with 1.0MKC1 (Thomas 1982).
The metals Na, K, Mg, Ca, Fe, Mn, Cu and Zn
were extracted for two hours using 2.0M HC1
(1:20 w/v), and its concentration in the ex-
tracts were measured by atomic absorption spec-
trophotometry. The extractant was randomly
chosen but was expected to remove the relative-
ly more soluble fractions of each metal, inclu-
ding the exchangeable and a large proportion
of the oxide and organic bound fractions. The
amount obtained is presented as HC1 - ex-
tractable metal.
RESULTS AND DISCUSSION
The chemical characteristics of the soil samples
are shown in Table 1. Each soil profile consists
of five distinct horizons, namely A, E, Bh, Bh/
s and C horizons. All samples show distinct
differences in chemical characteristics between
each horizons, particularly between the
bleached (E) and the spodic (Bh) horizons.
The pH distribution of each profile fol-
lows a similar general trend. It increases to a
maximum at horizon E, decreases to a mini-
mum at horizon Bh and then gradually increases
again at the lower horizons. Unlike ph, per-
cent organic carbon decreases significantly from
0.51 - 0.97% in horizon A to 6.05 - Q.()iY< in
horizon E. It, however, increases significantly
to 1.18-1.79% in horizon Bh before decreasing
gradually in the lower horizons. The lowest and
highest values obtained for horizon E and r>h
are in agreement with the observed bleached
and dark colour of the respective horizons. Ibis
contrasting levels of organic carbon between
the two horizons has been reported to be due
to mobilization of organic matter from the up-
per A horizons during intensive weathering and
leaching and later deposition in horizon Bh
(McKeague et al 1985; Fitzpatrick 1982;
Bridge 1982; Russell 1973).
Changes of cation exchange capacity
(CEC) down the soil profile at all sites also
follow a similar trend as that of organic carbon.
The high and low CEfc of horizons Bh and E
respectively is likely to be related to the signifi-
cant difference in organic carbon content of
the horizons since the organic component of a
soil is a well known contributor to soil's CEC.
The relationship between the CEC of soils and

TABLE 1
Some chemical characteristics of the respective horizons of the pod/ol samples PI, P2 and P>

PI P2 P3
Parameter A E Bh Bhs c; A E Bh Bhs C E Bh Bhs C

pH (H20) 4.8 5.8 3.9 4.3 4.0 4.6 5.1 4.2 4.3 4.3 4.2 4.7 4.0 4.1 4.4
% OrganioC 0.51 0.06 1.79 0.20 0.15 0.95 0.06 1.18 0.61 0.15 0.97 0.05 1.26 1.05 0.33
Exch. Cations
(me/lOOg)
Na 0.46 0.16 0.51 0.36 0.17 0.39 0.14 0.46 0.33 0.15 0.30 0.16 0.41 0.22 0.18
K 0.04 0.02 0.05 0.04 0.03 0.05 0.01 0.03 0.03 0.05 0.03 0.02 0.07 0.04 0.03
Ca 0.08 0.02 0.06 0.04 0.04 0.08 0.02 0.06 0.06 0.02 0.08 0.02 0.06 0,06 0.04
Mg 0.06 0.01 0.06 0.04 0.04 0.06 0.01 0.06 0.04 0.02 0.06 0.01 0.06 0.06 0.04
H 0.06 0.03 0.53 0.43 0.40 0.11 0.03 0.39 0.29 0.39 0.11 0.02 0.52 1.04 0.71
Al 0.03 0.03 0.39 0.27 O.(>4 0.03 0.01 0.27 0.27 0.33 0.03 0.03 0.42 0.51 0.48

60 PERTANIKA VOL. 12 NO. 1, 1989

 CHEMICAL CHARACTERISATION OF PODZOLS IN SABAH, EAST MALAYSIA

Amount e x t r a c t e d
100 200 300

<

§
z&
o

Amount extrac ted Amount e x t r a c t e d

100 200 'tOO 400 500 0.2 0.4 0.6 0.8 1.0

(b)

Mn

Amount e x t r a c t e d
1.0 2.0
•

(d) (e)
w W
c o
o x:
N CQ x;

Si C I is
CO
Cu I/ /
u u

i: HCI-extractable basic cations, Fe, Mn, Cu and Zn (in ug/g)

for podzol samples PI fo— oj, P2 (O — o ; and PS *3 <D—nj

PERTANIKA VOL. 12 NO. 1, 1989 61

 MARCUS JOPONY AND TAN YING CHIU
its organic carbon content is well described else-
where (Brady 1984; Stevenson 1982; Sch-
nitzer and Khan 1972). Differences in mine-
ralogical characteristics could also be a possible
factor. The amount of each individual ex-
changeable basic cation is less than 0.5 me/
lOOg and is generally of the order Na » Ca Mg
K. Exchangeable acidic cations meanwhile is of
the order H > AL. The distribution of these
exchangeable cations within the profile follow
a similar general pattern as that of the organic
carbon, and CEC significant differences in levels
of exchangeable H and AL are observed for
horizons E and Bh. This could be a possible ex-
planation for the higher acidity of horizon Bh
compared to horizon E. Apart from direct con-
tribution from the exchangeable H, the ex-
changeable AL is capable of undergoing hy-
drolysis in soil solution to produce hydrogen
ions (Brady 1984; Russell 1973; Sanchez
1972). The net result could therefore be a high
concentration of hydrogen ions in the soil
solution and thus higher acidity.
The amount of HCl-extractable Na, K, Mg
and Ca is relatively higher than the amount
exchangeable but follow a similar distribution
pattern within the soil profile. A plot of the sum
of HCl-extractable Na, K, Ca and Mg against
soil horizon is shown in Figure 1, clearly indicat-
ing a significant difference between horizon E
and Bh with respect to HC1- extractable basic
cations. A similar distribution pattern for HCl-
extractable Fe, Mn, Zn and Cu was also observ-
ed for all podzol samples (Figure 1). These re-
sults generally indicate that cations or metals
have been depleted from horizon E and accu-
mulated in horizon Bh.
Various hypothesis have been proposed
for the distinct differences in chemical charac-
teristics, particularly in metal and organic
carbon contents between horizon E and hori-
zon Bh. One such hypothesis was related to
organic matter (McKeague et al 1985). Sol-
uble humic substances, particularly fulvic acid,
produced by microbial attack on plant litter are
leached down the soil profile as rain percolate.
Due to its acidic nature, these organic substances
are thought to be capable of causing the break-
down of soil minerals (including clay minerals)
by disrupting the mineral structure and releas-
62 PERTANIKA VOL. 12 NO. 1, 1989
ing the component elements. The released ele-
ments then form soluble complexes (chelates)
with the humic subtance and are removed from
the surface horizons as the solutions percolate
downwards. The humic substance could com-
plex with additional metal cations as it perco-
late downwards and becomes increasingly satu-
rated with metal cations. Its solubility decreases
with increasing metal saturation and finally be-
comes insoluble and is precipitated as it reaches
horizon Bh. As a result, deposition of organic
matter as well as metal cations take place in
horizon Bh after being leached down from the
upper horizons.
CONCLUSION
This study showes that the podzol samples have
profile chemical characteristics similar to most
of the podzols described in the literature.
Marked maxima of organic carbon, cation ex-
change capacity, exchangeable cations and
extractable Na, K, Ca, Mg, Fe, Mn, Cu and Zn
(in this case, extracted with 2.0m HC1) occur
in the Bh horizon relative to the other hori-
zons.
ACKNOWLEDGEMENT
The authors wish to extend their thanks to Mrs.
Elizabeth Ginsos for typing this manuscript, and
to UKMS and the Lab technicians of Jabatan
Kimia FSSA for their technical support.
REFERENCES
ACRES, B.D., R.P BOWER, P.A. BURROUGH, CJ. FOI-
LAND, M.S. KALSI, P. THOMAS and P.S.
WRIGHT. 1975. The Soils of Sabah. Land Re-
Sources study 20. MOD England.
ANDRIESSE, J.P. 1969. A Study of the Environment
and Characteristics of Tropical Podzols in Sara-
wak (East Malaysia). Geoderma 2: 210-227.
BRADY, N.C. 1984. The Nature and Properties of Soils.
London: Macmillan.
BRIDGES, E.M. 1982. World Soils. Cambridge: Cam-
bridge University Press.
BURNHAM, C.P. 1984. The Forest Environment
Soils. In Tropical Rainforests of the Far-East. Whit-
more, J.C. (ed.). Oxford: Clarendron Press.
BURNHAM, C.P. 1968. Landscape and Soils in
Malaya. Malay. Agric. 7: 64-69.
DUCHAUFOUR, P. 1982. Pedology: Pedogenesis and
classifications (English Translation). London:
George Alklen and Unwin.
 CHEMICAL CHARACTERISATION OF PODZOLS IN SABAH, EAST MA1AYSIA
EVANS, LJ. 1982. Characteristics of Some Loamy
Textured Podzols in North Eastern Ontario. Can.
J. Sod Set. 62: 281-290.
FIT/PATRICK, E.A. 1983. Soil Their Formation Classi-
fication and Distribution. London: Longman.
FOTH, H.D. 1978. Fundamentals of Soil Science. New
York: John Wiley and Sons.
KLINGE, H. 1965. Podzol Soils in the Amazon
Basin./ Soil Sri. 16: 95-103.
MCKEAGUE, J.A., F. DECONICK, D.P. FRANZ
MEIER. 1985. Spodosols. In Pedogenesis and Soil
Taxonomy. II. The Soil Orders. Wilding, L.P., Smeck,
N.E., Hall, G.F. (eds.). Amsterdam: Elsevier Sci-
ence Publishers.
MOHR, E.C.J., F.A. VAN BAREN and J. VAN SCHUYLEN-
BORC;H. 1972. Tropical Soils. A Comprehensive
Study of their Genesis. Hague: Mouton-Ichtiar Baru
- Van Hoeve.
MuiR, A. 1961. The Podzols and Podzolic Soils.
Adv. Argon. 13: 1-56.
NELSON, D.W. and SOMMERS, L.E. 1982. Total Car-
bon Organic Carbon, and Organic Matter. In Me-
thod of Soil Analyses. Part 2. Page, A. L. Miller, R.H.
and Keeney D.R. (eds), Madison, Wisconsin USA.
PERTANIKA VOL. 12 NO. 1, 1989
RUSSELL, E.W. 1973. Soil Condition and Plant
Growth. London: Longman.
SANCHEZ, P.A. 1976. Properties and Management of
Soil in the Tropics. New York: John Wiley and Sons.
SCHNITZER, M. and S.U. KHAN. 1972. Humic Sub-
stances in the Environment. New York: Dekker.
STEVENSON, FJ. 1982. Humus Chemistry. Genisis,
Composition and Reaction. New York: John Wiley
and Sons.
THOMAS, G.W. 1982. Exchangeable Cations. In
Methods of Soil Analysis. Part 2: Page, A.L., Miller,
R.H. and Keeney, D.R. (eds.). American Society
of Agronomy, Inc. madison, Wisconsin USA.
TORRENT, J. and F. GOMEZ-MARTIN. 1985. Incipient
Podzolization Processes in Humic Acrisols of
Southern S p a i n . / Soil Sri. 36: 389-399.
TOWNSEND, W.N. 1977. Introduction to the Scientific
Study of the Soil. London: Edward Arnold Ltd.
(Received 20 November, 1988)
63
Pertanika 12(1), 65-70 (1989)

Changes in Surface Potential of Activated Carbon

Due to Adsorption of Ions
M. BADRI and KAREN A. GROUSE
Department of Chemistry
Faculty of Science and Environmental Studies
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

Key words: Activated carbon; pores; electrical double layer; adsorption; adsorption potential.

ABSTRAK
Pembinaan satu elektrod karbon dan sukatan ke atas pertukaran keupayaan di permukaan yang diakibatkan
daripada penjerapan ion dibincangkan sebagaifungsi kepada penjerapan dan pengionan kumpulan-kumpulan
berfungsi di permukaan karbon.

ABSTRACT
The construction of a carbon electrode and the measurement of the changes in the potential that developed
on its surface as a result of adsorption of ions are discussed as a function of the cation adsorption and ionization
of the functional groups on the surface of carbon.

INTRODUCTION the pores (Badri and Crouse 1987). These

It has been proposed that activated carbon
possesses many kinds of organic functional
groups lined up along the wall of the pores
(Hassler 1974). These functional groups react
with water to produce hydrogen ions which can
act as additional charge carriers in carbon under
an applied voltage (Badri et al 1984). As a
result of the reaction, the walls of the pores
would be lined with the negatively charged
conjugates of the acids. Since these conjugates
are immobile, the hydrogen ions would remain
in their vicinity in the pores thereby establish-
ing a charged double layer.
Activated carbon is also known for its
ability to adsorb cations. When cation adsorp-
tion takes place, these adsorbed species will
be fixed at the active centres on the walls of the
pores leaving the anions lined up alongside
them in the solution. Thus another charged
double layer of opposite polarity to the one
described above will be formed.
It was found possible to vary the degrees
of ionization and adsorption by changing the
concentration of the H+ on the solution side of
changes were reflected in the changes in the
potential developed across the double layer.
In this paper the effect of the adsorption of
some other cations will be described.
MATERIALS AND METHODS
All reagents were of analytical grade unless
otherwise specified.
Large pieces of activated carbon were pre-
pared by a method described earlier (British
Patent 1984). Samples thus prepared have very
low electrical conductivity, of the order of 1 x
10^ (ohm - cm) 1 (Badri et al. 1987). Fairly flat
pieces were chosen and were ground with fine
sand paper into discs of about 6 mm diameter
and 1.5 mm thickness. These discs were re-
fluxed in 1 M HC1 for about one week then con-
tinously washed with distilled water in a Soxhlet
apparatus for another week. After drying in an
oven at 110°C for 12 hours, two discs were glued
together with silver-loaded epoxy resin and
then sealed in glass tubing of appropriate
diameter with Araldite epoxy resin. After drying
at room temperature, the glass tube was filled
 M. BADRI AND KAREN A. GROUSE
with -0.5 ml 1M HC1. An Ag/AgCl electrode
(prepared according to Shoemaker and Gar-
land) was inserted into the tube. This assembly
is henceforth referred to as the carbon elec-
trode. For the present experiments, three such
electrodes were used.
Solutions of a number of chlorides of vari-
ous concentrations were prepared. These were
kept in small, covered polyethylene bottles and
were equilibrated at 25 ± 0.2°C. The potential
of the double layer which developed when the
carbon electrode was dipped into the prepared
solution was measured against a Coleman calo-
mel electrode using a dc microvoltmeter. The
voltmeter reading was recorded when a cons-
tant value was registered. After each measure-
ment, both electrodes were rinsed with distilled
water and gently dried with soft tissue paper
before proceeding to a solution of different
concentration. In addition, the carbon elec-
trodes were also soaked in distilled water for 16
hours before being used again for measure-
ments with a test solution containing a differ-
ent cation.
RESULTS
The potentials of the double layer measured
are given in Table 1 and are plotted as a function
of -log concentration, pC, in Ftgs 1 to 4.
Fig. 1 shows this relationship for the va-
lues obtained from Electrode 1 in HC1 and
NaCl solutions. Reproducibility was good with
an average uncertainity of ± 4rnV.
The rate of change of the potential in the
whole range of pC from 0 to 7 was found to be
not linear. To simplify discussion, each of the
curves given in Fig. 1 is separated into two re-
gions. Region 1 refers to that for pC range from
0 to 3 (or 4) where the rate of change was high
and almost linear except for a slight shoulder
at pC - 1 . Region 2 refers to that for higher pC
range where the rate of change was lower and
in which maxima and minima were interspersed.
Similar potential-concentration relation-
ships were also found for other salts, Figs. 2 to
4. The first set of values, Fig. 2(A), was obtained
from Electrode 2 with solutions of HC1, NaCl,
KC1 and LiCl when measurements were done
in that order.*The potentials obtained for KC1
and LiCl solutions in repeat measurements
66 PERTANIKAVOL. 12 NO. 1, 1989
after the first set was completed are shown in
Fig. 2(B) to illustrate the possible error that
might be introduced in a potential measure-
ment if inadequate desorption of adsorbed spe-
cies had not taken place. Curve 1 was obtained
immediately after the first set while Curves 2
were obtained after soaking Electrode 2 in dis-
tilled water for approximately 12 hours. Elec-
trode 2 was soaked in HC1 (6M) for 3 days and
then in distilled water for three days in an
attempt to restore it. The results which are
shown in Fig. 3 seem to indicate, some improve-
ment. However, the original characteristics of
Electrode 2 were not fully restored. The linear-
ity of the rate in the pC range of 0 to 1 was lost
and the potential measured for 1M solution
was found to have increased from the original
value of — 15mV to ~40rnV. Data obtained for
the chloride of Ca2+(aq) are also included in
Fig. 3 for comparison.
The potentials obtained from Electrode 3
for the solutions of doubly- and triply -charged
cations are given in Fig. 4. The values obtained
for KC1 solutions are included for the purpose
of comparison. The magnitude of charge
present on the cations does not appear to affect
the potential characteristics of the double layer.
DISCUSSION
Activated carbon is known to have a variety of
active centres which are either basic or acidic.
The acidic groups such as carboxylic and phe-
nolic may dissociate in water to give H+ depend-
ing on the pH of the solution. The basic groups
can be carbonyl, ether, quinone, benzpyrene
and so on. Both groups are responsible for
the adsorptive character of activated carbon.
In these experiments, the opposite sur-
faces of the carbon discs were in contact with
solutions of different concentrations. Since the
concentration of the HC1 solution that was in
contact with the internal surface of the disc was
kept constant, the potential gradient developed
across the charged double layers due to ioniza-
tion of the functional groups and adsorption of
H+ would remain constant as well. Thus changes
in the potential observed as the external sur-
face of the disc was dipped into the test solu-
tions must be due to the changes in the poten-
tial gradient across the charged double layer
of the external surface alone.
 CHANGES IN SURFACE POTENTIAL OF ACTIVATED CARBON DUE TO ADSORPTION OF IONS

TABLE 1
Double layer potential* of activated carbon

Electrode Solution -log Concentration, pC

0 1 2 3 4 5 6 7

1 HC1 20 57 87 113 103 110 129 110

HC1 20 57 83 109 107 111 120 128
NaCl 49 69 84 101 120 122 132 110
NaCl 43 62 79 97 119 120 135 100
2 HC1 -18 40 76 102 94 102 117 100
(1st set) NaCl -17 19 55 88 120 113 118 103
KCl -9 27 67 101 110 123 120 110
LiCl -15 20 61 93 120 131 138 89
KCl (Curve 1) 61 69 87 100 96 100 86 94
KCl (Curve 2) 51 67 90 113 122 128 130 119
LiCl 67 70 90 106 122 123 129 118
(2nd set) NaCl 40 65 90 108 122 122 118 111
NH4C1 58 65 86 103 119 117 127 123
KCl 51 67 90 113 122 128 130 119
CaClt> 44 65 90 108 121 118 139 120
LiCl 67 70 90 106 122 123 129 115
3 KCl -27 23 56 73 69 74 76 75
A1C1, -42 -1 39 72 82 74 75 70
MgC2 -42 0 39 68 72 75 76 64
CaCl, -42 -2 35 63 78 75 76 67

* All potentials are expressed in mV.

With respect to this, one might suggest Since the adsorbed ions were immobilized,
that the changes in the potential were due to their negatively charged conjugates, X, would
the ohmic drop across the electrode as the be lined up along the surface but on the solu-
solution filled the pores. We have discounted tion side of the pores. An array of these oppos-
this theory because according to our previous ing charges would give rise to an electrical
work (Badri et al. 1985) the conductivity of double layer on each of the opposite surfaces
activated carbon soaked in HC1 followed the of the disc. We denote the respectve potentials
Onsager equation in the pH range of 1.6 to 6.7 as <x. for the internal surface and c^ for the
whereas the present results do not. external surface.
As the surface of the disc comes into The acidic functional groups on the
contact with a solution, adsorption occurs at surface could also undergo ionization when the
the basic active centres, since these electron- pores are flooded with the solutions. Taking
rich groups would share the electron clouds the phenolic group as an example, the ioniza-
with the adsorbed species. As a result of this, a tion reaction of these acidic groups could be
shift of the electron cloud from the graphite written as
rings of the carbon crystallite to the surface (surface) (solution) (solid) (surface) (solution)
would occur. Taking the carbonyl as an exam-
ple, the reaction could be written as: nfc OH + oq - * H+(oq)

(solid) (surface) (solution) (solid) (surface) (solution) It is assumed that most of these acidic
groups were weakly ionized and their popula-
tion was relatively small compared to the total
number of adsorption sites. However, when

PERTANIKA VOL. 12 NO. 1, 1989 67

 M. BADRI AND KAREN A. CROUSE

140

120

120

100

1
< 80
t-
z
UJ

60

20-f

2 3 4 5 b /
I 2 5 4 5 6
-LOG CONCENTRATION, pC
-LOG CONCENTRATION, pC

Fig. 2. Adsorption potential of activated carbon

Fig. 1. Double layer potential of activated carbon Electrode 1 in Electrode 2. (A) 1st set of readings
HCl and NaCl solutions (B) 2nd set of readings

2 3 4 5 6 2 3 4 5 6
-LOG CONCENTRATION, pC -LOG CONCENTRATION, pC

Fig. 4. Adsorption potential of activated carbon

Fig. 3. Adsorption potential of activated carbon Electrode 3 in solutions of doubly-and triply-
Electrode 2 after partial regeneration of active sites charged cations

68 PERTANIKA VOL. 12 NO. 1, 1989

 CHANGES IN SURFACE POTENTIAL OF ACTIVATED CARBON DUE TO ADSORPTION OF IONS
ionization took place, an array of negatively
charged conjugates would line up the surface
of the pores with protons staying near them
on the solution side. This situation would also
give rise to an electrical double layer but the
potential gradient, p, across the layers would be
small and opposite to that discussed above for
the adsorption.
However, such ionization of the acidic
groups might further increase the activity of
other electron-rich groups at least by induction
if not by outright transfer of the electron cloud
through the graphite rings of the carbon crys-
tallite. In this way the less active groups such as
peroxides or etheral oxygen which may be
present on different rings of the same graphite
layer would be transformed into more active
sites and hence increase the adsorptive capacity
of the surface. An example of such a mecha-
nism could be given as follows:
o,
T As the concentration was reduced, dissociation
A strong acid, such as HC1, at a concen-
tration of 1M, would supress the ionization of
the acidic groups. Thus, being weak and rela-
tively small in number, the acidic groups would
contribute very little to the development of the
potential at the internal surface of the disc.
Hence, the potential gradient at the internal
surface of the disc could be safely assumed to
be due to adsorption, a. However, at the exter-
nal surface of the disc this would be true only
at low pC, and therefore, the potential gradient
was determined by both av and P^, the potential
due to the ionization, Pt> being very small
compared to a^. Thus the net potential, V, that
was measured was the sum of a., ctt> and pe:
V = a - (a, - pti)
In region 1, where the concentrations of
the cations were large, extensive adsorption
occured resulting in a large at. and small V. At
lower concentrations of the cation, V increased
because c*e decreased. The trend continued until
a maximum was reached at pC = ~3 or 4 where
the effect of the reversal potential due to
ionization took place. It is assumed that at these
pCs, some of the acidic groups began to ionize.
PERTANIKAVOL. 12 NO. 1, 1989
On ionization, the activity of other electron-
rich functional groups was enhanced according
to I above, thus increasing the adsorption. This
resulted in greater oit which in turn reduced V.
The subsequent increase in V could per-
haps be explained as the direct result of com-
pletion of ionization of these weak acids. Once
ionization was complete, there would be no
more less-active sites to be promoted accord-
ing to Scheme I and therefore a decrease in the
concentration of the action would decrease at.,
and hence V would increase again. The subse-
quent increase of V in region 2 could also be
due to the dissociation of a protonated pyrone
type structure which was formed according to
scheme II (Parentich and Kinsella, 1984).
ii
The protonated species contribute to 0te,
whereas the unprotonated species is neutral.
of the protonated species took place hence
reducing a.
Similarly, if a benzpyrene group which is
found in many natural products is present
(Tsuchida and Muir 1986), the surface of
activated carbon would also have species such
as III.
,
III
which contribute to the development of a. At
lower concentration of acids, the charged spe-
cies would revert to its neutral form according
to scheme IV
rV- OH
IV
and at the same time reduce otc to increase V.
The fact that there were three minima in
a V-pC curve (including the shoulder in Region
1) might indicate that there were at least three
acidic groups, each comprising members hav-
69
 M. BADRI AND KAREN A. CROUSE
ing similar ionization constants. The identifica-
tion of these species may be determined only
through further experiments.
There seems to be very little difference in
the effect of adsorption of various cations on
the adsorption potential of the carbon surface.
Adsorption of cations carrying higher charges
would be expected to suppress the activity of
the adjacent sites more than singly-charged
cations. If this did not occur, equal number of
Ca2+ adsorbed, for example, would decrease V
by 2-fold over that due to the adsorption of
K\ But the fact that at pC = O, V was just slightly
smaller for the doubly-charged cations indicates
that the number of occupied sites was only
fractionally greater that one-half of the total
sites occupied by the singly-charged cations.
After adsorption, the characteristics of the
carbon electrode could be reestablished simply
by soaking it in distilled water except for the
case where the adsorbed cation was Li+. Li4 was
more strongly adsorbed compared to other
cations including AP. This is probably due to
its ability to enter the micropores of the acti-
vated carbon. This further substantiates the con-
clusion that the magnitude of the charge on a
cation did not play a main role in determining
the strength of the bond formed at the adsorp-
tion sites. It is interesting to note, however, that
a surface contaminated with some preadsorbed
species gave higher values of V, Fig. 2. It seems
that pre-adsorption did not shift the entire
ordinate of Fig. 2 towards a higher value but
affected only the adsorption side of the V-pC
•
70 PERTANIKA VOL. 12 NO. 1, 1989
curve, i.e. Region I. This could only mean that,
if the solution were to contain various cations,
these cations would compete for the same ad-
sorption sites.
REFERENCES
BADRI, M., GROUSE, K.A. and HARUN AWANG. 1984.
Conductance Measurement of Water-Sorbed
Activated Carbon. Pertanika 7 (3): 59-65.
BADRI, M., CROUSE, K.A. and HARUN AWANG. 1985.
The Effects of Sorbed Alkali Metal Ions on the
Conductance of Activated Carbon. Pertanika 8 (3):
337-341.
BADRI, M. and GROUSE, K.A, 1987. Adsorption
Potential of Activated Carbon. In Proc Asian Science
6* Tech. Congress *87, Kuala Lumpur, Malaysia, 14-
17 Oct. 1987. (In press).
HASSLER, K.W. 1974. Purification xvith Activated Car-
bon. Industrial, commercial and Environmental. U.S.A.:
Int. Textbook Co.
PARENTICH, A. and KINSELLA, B. 1984. Titrimetric
Studies on Activated Carbons. In Aus. InsL of
Mining and Metallurgy Regional Conference.
SHOEMAKER D.P. and GARIAND, C.W. 1962. Experi-
ments in Physical Chemistry, p. 445. New York:
McGraw-Hill.
TSUCH1DA, N. and MuiR, D.M. 1986. Studies
on Role of Oxygen in the Adsorption of Au
(CN)V and Ag(CN) 2 onto Activated Carbon.
Metallurgical Trans. B 17B 529-533.
UNIVERSITI PERTANIAN MALAYSIA. 1984. British Pa-
tent No. 2086864.
(Received 17 September, 1988)
.
Pertanika 12(1), 71-78 (1989)

A Simple and Clean Method for Methoxymethylation of Phenols

FAUJAN B, H. AHMAD and J. MALCOLM BRUCE1
Department of Chemistry,
Faculty of Science and Environmental Studies,
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

Key words: Methoxymethyl ethers of phenols.

ABSTRAK
Satu kaedah mudah dan bersih, untuk penyediaan metoksimetil eter (MOM = CH2 OMe) bagi fenol yang
membawa ikatan hidrogen kumpulan hidroksi dalam molekulnya akan dibincangkan. Tindakbalas 2,5-
dihidroksibenzaldehid (1) dengan metoksimetilklorida - metil asetat dalam pelarut eterpada suhu biliky meng-
hasilkan 57% 2,5- bis(metoksimetoksi) benzaldehid (2). Dalam keadaan tindakbalas yang sama 75%
metoksimetil eter (6) telah dihasilkan daripada salisilaldehid. Penghasilan sebanyak 61-81 %, tidak dibaiki,
bagi metoksimetil eter untuk beberapa fenol yang tidak mempunyai ikatan hidrogen di alas telah juga
dihasilkan.

ABSTRACT
A simple and clean procedure for the preparation of methoxymethyl ethers (MOM = CH2 OMe) of phenols having
internally hydrogen bonded hydroxy groups is described. Thus treatment of 2,5-dihydroxybenzaldehyde (1) with
a 1:1 mixture of methoxymethyl chloride-methyl acetate in ether at room temperature gives 2.5-bis-
(methoxymethoxy)benzaldehyde (2) in 57% yield; under similar conditions, the methoxymethyl ether (6) of
salicylaldehyde was isolated in 75% yield. Yields of 61-81 %, not optimised, of methoxymethyl ethers of several
phenols lacking internal hydrogen bonding were also obtained.

INTRODUCTION Several alternative methods for methoxy-

The methoxymethyl ether moiety is a useful
hydroxy protecting group for phenols, alcohols,
and carboxylic acids. Methoxymethylation is
sometimes superior to tetrahydropyranylation,
since the latter results in the formation of new
assymmetric center(s); with diols and optically
active alcohols, a mixture of diastereomers is
formed, complicating both purification and
spectroscopic analysis (Fuji et al 1975). Prepa-
rations of methoxymethyl ethers are based
mostly on the reaction of a phenoxide anion
with methoxymethyl chloride (Greene 1981).
However, such a procedure was not suitable for
our purpose, the preparation of bis(metho-
xymethoxy)benzaldehyde (2) from 2,5-dihy-
droxybenzaldehyde (1).
methylation which avoid the use of methoxy-
methyl chloride present some difficulties. The
use of methylal and a large molar excess of
phosphorus pentoxide (Fuji etal 1975) causes
difficulties in work-up, particularly of methoxy-
methyl ethers of small molecular weight. Based
on Fuji's procedure, Yardley and Fletcher,
(1976) reported that 3.5 g of (3) required 85
g of phosphorus pentoxide and a final neutrali-
sation volume of 4 litres. They then reported
on the use of methylal and 4-toluenesulfonic
acid in the presence of molecular sieves (to
remove methanol) to facilitate the preparation
of some methoxymethyl ethers. However, their
procedure failed to afford either the methoxy-
methyl ether of 2-acetylphenol (4) or the bis-

1
Department of Chemistry, University of Manchester, Manchester, Ml3 9PL, England.
 FAUJAN B. H. AHMAD AND J. MALCOLM BRUCE
methoxymethyl ether of 2,2-dihydroxybenzo-
phenone (5). The difficulty may be due to the
internally hydrogen bonded hydroxy groups in
these compounds. Recently, the use of methylal
and phosphorus oxychloride in toluene at 65°C
was reported to give the methoxymethyl ether
(6) of salicylaldehyde in 90% yield (Sch-
outen 1985).
We herein report a clean and simple pre-
paration of methoxymethyl ethers, particularly
from substrates having internally hydrogen
bonded hydroxy groups, such as that in alde-
hyde (1), which illustrates the importance of
correct choice of a solvent. The procedure was
found to be superior to that generally used.
MATERIALS AND METHODS
Proton magnetic resonance spectra, in p.p.m.
with respect to internal tetramethylsilane, were
measured on a Perkin-Elmer R34 instrument
at 220 MHz, and a Varian SC300 instrument at
300 MHz as stated. Coupling constants for the
aromatic protons were in the normal ranges.
Resonances assigned to hydroxyl groups were
removed by addition of D,O.
Mass spectra were recorded on Kratos
MS25 and MS30 instruments. Melting points
were recorded on a Kofler block and were un-
corrected.
Infrared spectra were recorded on a
Perkin-Elmer FTIR 1710 spectrometer as Nujol
mulls, films or solutions as stated.
Methoxymethylation of Phenolic Hydroxy
Groups: A General Procedure for Preparation of
Compounds (13 a-i) and (14 a-f).
To a stirred solution of the hydroxy com-
pound (hydroxybenzene, hydroxyaldehyde,
hydroxyketone, or hydroxycarboxylic acid^ (1.0
mmole) in ether (5 ml) (Note 1) under a nitro-
gen atmosphere was added methoxymethyl
chloride (1.5 mmole), as a 1:1 mixture with
methyl acetate (Note 2), and triethylamine (2.0
mmole) (Note 3). The mixture was stirred at
room temperature for about 24 h. and the white
precipitate was then removed by filtration.
Removal of the solvent gave the methoxymethyl
ether, usually as a liquid, which was purified
1-\Uthoxymethoxy-2<yclohexen-lHone was obtained in 75% yield from the corresponding hydroxy compound on
treatment with methoxymethyl chloride-methvl acetate in the presence of triethylamine in dichloromethane at 0°C
72 PERTANIKA VOL. 12 NO. 1, 1989
either by washing with aqueous 5% sodium
hydroxide or by distillation.
Note L Ether (5 ml) was used as the sol-
vent for every 1.0 mmole of the hydroxy com-
pound, except for those hydroxy compounds
which were not very soluble in ether when
more ether was used.
Note 2. Methoxymethyl chloride (1.5
mmole) (Amato et al 1979) was used for each
hydroxy group present in the starting material.
Note 5. Triethylamine (2.0 mmole) was
used for every 1.5 mmole of methoxymethyl
chloride used in the reaction mixture. Excess
of amine ensured that the reaction mixture
remained basic throughout.
The compounds prepared are listed in
' Tables 1 and 2. Their analytical and spectros-
topic data are shown in* Tables 3 and 4, respec-
tively.
RESULTS AND DISCUSSION
In connection with our interest (Ahmad and
Bruce, 1986) in developing a new synthetic
route to the aglycones of the anticancer anthra-
cyclines, we required the hydroxy protected
aldehyde (2). However, treatment of 2,5-dihv-
droxybenzaldehyde (1) with a 1:1 mixture of
methoxymethyl chloride-methyl acetate (Amato
et al, 1979) in dichloromethane* (Khan and
Bruce 1985) in the presence of triethylamine,
either at room temperature or at reflux, gave
only 5% of the desired aldehyde (2), the major
product being the mono-methoxymethyl ether
(7). Similar reactions using pyridine as the base
in either dichloromethane, tetrahydrofuran or
ether failed to give the desired aldehyde (2):
only starting material (1) was isolated. Attempt-
ed methoxymethylation of aldehyde (1) in the
presence of powdered 4A molecular sieves to
absorb hydrogen chloride (c,f. Yardley and
Fletcher 1975) again gave the mono-methoxy-
methylation product (7). The difficulty in pre-
paration of (2) may be due to internal hydro-
gen bonding [as (7a)] in the starting material
(1).
1,4-Bis(methoxymethoxy) benzene (8) has
previously been obtained by heating hydroqui-
 A SIMPLE AND CLEAN METHOD FOR METHOXYMETHYIATION OF PHENOLS

OMOM QH
CHO .CHO

CHO

(3)

H
COMe

(4)

H
9 b
CHO

(6) OH

(7a)

MOM =r CH 2 OMe

PERTANIKA VOL. 12 NO. 1, 1989 73

 FAUJAN B. H. AHMAD AND J. MALCOLM BRUCE

TABLE 1
Methoxymethyl (MOM) ethers of some 1,2,4-trisubstituted benzenes 13 (a) .

Isolated, b.p. ( ° C / m m H g )
(13) Yield (%)

(a) Rl = R2 = MOM, R3 = H 61 76-80/0. l ( b )

(b) R1 = R2 = MOM, R* = OMe 76 80-86/0.1
(c) Rl = R2 - MOM, K" = CHO 57 56-60/0.05
(d) R1 = R2 = MOM, R:1 - COPh 23 9.6-100/0/0.1
(e) R1 - Rs = MOM, R3 = COjMe 18 Not determined
(0 R' = R2 = MOM, Rs = CO2MOM 17 Not determined
(g) R1 = H, R2 = MOM, Rs = CHO 80<c) 50-56/0.1
(h) R1 = H, R2 = MOM, Rs - CO, H 82 [m.p 104-106°C]
Decomposed on
attempted sublimation
(i) R1 = H, R2 = MOM, R* = COMe 10 Not determined

(a>
Prepared from the corresponding hydroxy compounds. Except for entry (c), yields were «ot optimised.
(b)
Mamedov and Mamedova (1962), b.p. 136-137°C/5 mmHg.
(O
The compound was prepared in refluxing dichloromethane using powdered 4A molecular sieves.

TABLE 2
Methoxymethyl (MOM) ethers of some 1,2-disubstituted benzenes (14)(ii)

(14) Isolated b.p.(°C/mmHg)

Yield(%)

(a) R1 = Br; R2 = OMOM 81 60-64/0.1

(b) R1 * OMOM; R2 = H 75 (b) 60-66/0.1
(c) R1 = OMOM; R2 * OH 82 60-64/0. l (c)
(d) R1 = OMOM; R2 = OMe 16 Not determined (d)
(e) R1 - OMOM; R2 = OMOM 10 Not determined
(f) R1 = OMOM; R2 = Me 10 Not determined
u)
Prepared from the corresponding hydroxy compounds. Yields were not optimised.
(b)
This compound is known; prepared in 90% yield by treatment of the corresponding aldehyde with methylal and
phosphorus oxychloride in toluene at 65°C (Schouten, 1985).
(c)
Dunn and Bruice (1970), white solid, m.p. 63-64°C.
(d)
Dunn and Bruice (1970), b.p. 72-73°C/0.025 mmHg.

74 PERTANIKA VOL. 12 NO. 1, 1989

 A SIMPLE AND CLEAN METHOD FOR METHOXYMETHYIATION OF PHENOLS

TABLE 3
Characteristics of methoxymethyl ethers of some 1,2,4-trisubstituted benzenes (13).

Com- Elemental P m.r (220 MHz, CDC13)(a) I.r/cm 1

pound analysis
or M+ OMe OCH2 ArH Other (film)<b)
13(a) M-; 198.0892 3.42 5.04 6.92 (s,4H) 15000s
(s,6H) (s,4H)
13(b) C,57.6, 3.46 5.12 6.55(dd,lH) 1511m
H,7.2% 3.50 5.14 6.65(d,lH) 1153m
3.85 7.05(d,lH) 1009m
13(c) C.58.6; 3.46 5.16 7.02(d,lH) 0.28 1680s
H,6.3% 3.50 5.21 7.26(dd,lH) (s,CHO) 1490m
7.52(d,lH) 1385m
13(d) C68.1; 3.28 4.96 7.06(m,lH) 1669s
H,6.0% 3.46 5.13 7.04(s,lH) 1597m
7.05(s,lH) 1493s
7.44(m,2H)
7.53(m,2H)
7.86(d,2H)
lS(e) M+;256.0947 3.44 5.12 7.13(m,2H) 3.86 1720s
3.48 5.16 7.46(m,lH) (s,Co2 Me) 1490m
13(0 C.54.1; 3.47 5.02 7.08(s,lH) 1736s
H;6.4% 3.52 5.17 7.10(s,lH) 1498s
3.54 5.42 7.42(d,lH)
13(g) C59.6; 3.43 5.12 6.92(d,lH) 9.29 3100-
H,5.7% (s,CHO) 3600b
7.22(m,2H) 10.65 1660s
(s,OH)
13(h) C.54.9; 3.50 5.16 6.96(d,lH) 10.10 3100-
H,5.6% 3600b
6.96 7.25(dd,lH) (bs,20H) 1682s
7.60(d>lH) 1618s
1489m
13(i) M"196.0731 3.50 5.12 6.88(d,lH) 11.92 3150b
(s,OH) 1640s
7.80(dd,lH) 2.62
(s,COMe)
• 7.42(d,lH)
11
P.m.r. spectra of 13(c,e,j) were recorded at 300 MHz; those of 13(a,0 were recorded at 60 MHz. Signals due to OMe
and OCH., are singlets.
11
I.r. spectrum of I3(i) in CDCI^; of 13(h) in Nujot.

none in ether with methoxymethyl chloride and
dimethylaniline, in about 60% yield (Mamedov
and Mamedova 1962). In our hands, compound
(8) was more easily prepared by treatment of
hydroquinone with a 1:1 mixture of methoxy-
methyl chloride - methyl acetate in the pre-
sence of triethylamine at room temperature, in
ether, also in about 60% yield (Scheme 1). There-
fore the mono-methoxymethylation product (7),
which was obtained previously as described
above, was treated with a 1:1 mixture of methoxy-
methyl chloride-methyl acetate in the same man-

PERTANIKA VOL. 12 NO. 1, 1989 75

 FAUJAN B. H. AHMAD AND J. MALCOLM BRUCE

TABLE 4
Characteristics of methoxvmethyl ethers of some 1,2-disubstituted b e n z e n e s (14).

Com- Elemental P.m.r. (220 MHz, CDCl's)

(a)
I.r./cm 1
pound analysis OMe OCH2 ArH Other (film)

14(a) C,44.2;H,3.7; 3.50 5.42 7.29(m,2H) 1740s

Br, 32.4% 7.60(m,lH)
7.80(m,lH)
14(b) C,65.3;H,6.1% 3.42 5.22 7.02(t,lH) 10.42 1690s
(s,CHO) 1600s
7.16(d,lH)
7.48(td,lH)
7.78(dd,lH)
14(c) C,59.3;H,5.9% 3.56 5.51 6.90(t,lH) 8.66 3210-
(s,OH) 3004b
7.00(d,lH) 1630s
7!48(td,lH) 1615s
7.93(td,lH) 1486m
14(d) C,61.2;H,6.4% 3.54 5.29 7.08(td,lH) 3.90
(s,CO Me)
7.20(d,lH) 1731s
7.46(td,lH) 1755m
7.80(dd,lH)
14(e) C,59.0;H,6.4% 3.35 5.10 6.93(td,lH) 1734s
3.39 5.30 7.09(d,lH) 1602s
7.32(td,lH) 1488s
7.70(dd,lH)
14(f) C,66.5;H,6.7% 3.50 5.28 7.05(td,lH) 2.62
(s,COMe)
7.18(d,lH) 1677s
7.45(d,lH) 1598m
7.45(td,lH) 1483m
7.72(dd,lH) 1454m
(O
Signals due to OMe and OCH 2 are singlets.

OMOM
MeO .CHgC l/PhNJfe 2 /55 C / E t 2 0
or

MeO . N/20OC/Et20

Scheme 1

76 PERTANIKA VOL. 12 NO. 1, 1989

 A SIMPLE AND CLEAN METHOD FOR METHOXYMETHYIATION OF PHENOLS

ner as outlined for the preparation of (8): this
afforded the required bis(methoxymethoxy)-
benzaldehyde (2) in 60% yield. Hence, treat-
ment of 2,5-dihydroxybenzaldehyde (1) with 2-
3 mol of methoxymethyl chloride-methyl ace-
tate in the presence of triethylamine in ether
(ins-tead of dichloromethane as before), gave
the desired methoxymethyl ether (2) in 57%
yield. The latter route reduces to one step the
preparation of (2) from the corresponding
aldehyde (1) (Scheme 2). This procedure is
clean and simple, and illustrates the importance
of correct choice of solvent.
To our knowledge, the use of ether as
solvent for preparation of this type of methoxy-
methyl ether has not been previously reported
on. Therefore, it was of interest to explore the
use of the method for the preparation of other
methoxymethyl ethers, particularly from subs-
trates having internally hydrogen bonded
hydroxy groups similar to that in aldehyde (1).
Models of general structures (9) and (10) were
used. The progress of reaction was easily fol-
lowed by observing the formation of triethylam-
monium chloride which precipitated from
solution (Scheme 3).
Details of the methoxymethyl ethers which
were prepared are summarised in Tables 1 and
2. These show that the substrates without an
internal hydrogen bond gave 60-81% of the
corresponding methoxymethyl ethers ['a' and
'b' (Table 1) and V (Table 2)].Also, the
methoxymethyl ester (11)* was prepared from
the corresponding acid (12) in 81% yield. It is
worth noting that for the trisubstituted ben-
zenes (9) (Table 1), the yield of bismethoxy-
methyl ether decreases in the order R = H, Ph,
OMe, OH. In contrast, for the disubstituted
benzenes (10) the yield of bismethoxymethyl
ether decreases in the order R = OH, H, OMe,
Me. This order may be due to the solubility of
the starting materials. As expected, the mono
methoxymethyl ethers of the trisubstituted
benzenes (9) were isolated in high yield [en-
tries 'g' and 'h' (Table 1)]. In contrast, it was
difficult to prepare the bismethoxymethyl ether
of 2\5t-dihydroxyacetophenone: only its 5'-
monomethoxymethyl ether was obtained, in
10% yield (entry Y, Table 1).

CHO
MeO .CH2C l-MeC0 2 tte/ MeO . -KeC 0

Et
C/Bt 2 0
OMOM
(7)

OMOM
CHO
MeO . -MeC 0 2 K -CHO
Et N/20 C/Bt 2 0

OMOM
(2)
Scheme 2

Compound (11), oil, b.p. 100-106°C/0.1 mmHg: (Found M". 330.1103); C |8 H |B O 6 requires M. 330.1116.
It had 8 (22OMH/,CDC1,), 3.35(3H,s,OMe), 3.48(3H,s,OMe), 3.62(3H,s,OMe), 5.26(3H,s,OCH 2 ),6.85(lH,d,H-3'),
7.08(lH,dd,H-4), 7.32(lH,dd,H-3), 7.42(lH,d,H-6), 7.52(1 H,td,H-5), 7.58(lH,dt,H-4), 8.01(lH,dd,H-6); 8^ (film)
1658s, 1727s cm 1 ,

PERTANIKA VOL. 12 NO. 1, 1989 77

 FAUJAN B. H. AHMAD AND J. MALCOLM BRUCE

OH OMOM
Bt,N/Et 2 O
MeO-CHgCl

Scheme 3

OMe

(10) OMe

(9) (11* R = MOM

(12) R=r H

REFERENCES GREENE, T.W. 1981. Protective Groups in Organic Syn-

thesis. New York: Wiley-Interscience.
AHMAD, F.B.H. andJ.M. BRUCE. 1986. University
KHAN, AJ. and J.M. BRUCE. 1986. University of
of Manchester, unpublished work.
Manchester, personal communication.
AMATO, J.S., S. KARADY, M. SLETZINGER, and L.M.
MAMEDOV, Sh. and A,R. MAMEDOVA. 1962. Esters
WEINSTOCK. 1979. A New Preparation of Chlo-
romethyl Methyl Ether Free of Bis[chloromethyl of Glycols and Their Derivatives. XLI. Synthesis of
Ether. Synthesis 970-971. Alkoxy Derivatives of Methyl Ethers of Phenols,
7A Obshchei Khim. 37: 407-410; Chem. Abstr., 1963,
DUNN, B.M. and T.C. BRUICE. 1970. Steric and
58: 466a.
Electronic Effects on the Neighbouring General
Acid Catalyzed Hydrolysis of Methyl Phenyl Acetals S( HOI TEN, H.G. 1985. U.S. Pat. US 4,500,738;/
of Formaldehyde./ Am. Chem. Soc 92: 2410-2416. Synthetic Methods, 1985, 11: 76289A; Chem. Abstr.,
1985, 102: 184823/.
Fuji, K,, S. NAKANO, and E. FUJITA. 1975. An Im-
proved Method for Methoxymethylation of Alco- YARDLEY, J.P. and H. FLECTCHER. 1976. Introduc-
hols under Mild Acidic Conditions. Synthesis 276- tion of the Methoxymethyl Ether Protecting
277. Group. Synthesis 244,
(Received 27 June, 1988)

78 PERTANIKAVOL. 12 NO. 1, 1989

Pertanika 12(1), 79-82 (1989)

The Differences in Thermal Breakdown of H2PtCl6 and cis-

Pt(NH3)2Cl2 in Carbon Furnace Atomic Absorption.
MD. JELAS HARON and WAN MD. ZIN WAN YUNUS
Department of Chemistry,
Faculty of Science and Environmental Studies,
Universiti Pertanian Malaysia,
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia
•

Key words: Thermal decomposition; platinum compound; carbon furnace atomic absorption
spectrophotometry.

ABSTRAK
Kajian menggunakan analisis termogravimetri dan penyerakan x-ray menunjukkan larutan asid heksaklo-
roplatinik (H^tClJ, yangselalu digunakan sebagai larutan piawai di dalam teknik spektrofotometri serapan
atom (AAS), diubah kepada PtCl2 yang mudah merwap sebelum menjadi logam platinum. Sebaliknya larutan
cis-diaminodikloroplatinum (II) (Cis-Pt(NH J2Cl2; cis-DDP) diubah terus kepada logam platinum. Perbezaan
mekanisma penguraian oleh haba ini mungfan menjadi salah satu daripada sebab mengapa isyarat spektro-
fotometri serapan atom relau karbon (CFAAS) bagi H^tCl^ lebih rendah daripada cis-DDP.

ABSTRACT
A study using a combination of thermogravimetric and x-ray diffraction analyses has shown that a solution
of hexachloroplatinic acid (H^tClJ, usually used as a standard in atomic absorption spectrophotometric (AAS)
measurements, was broken down to yield volatile PtCl2 before being converted to metallic platinum. On the
other hand, a solution ofcis-diaminodichloroplatinum(II) (cis-Pt(NH')2Cl2; cis-DDP) was converted to metallic
platinum in a single step. The differences in their thermal decomposition mechanisms could be one of the reasons
as to why carbon furnace atomic absorption spectrophotometric (CFAAS) signal from H^tCl^ was found to
be lower than that of cis-DDP in aqueous solution.

INTRODUCTION 1975). The same explanation was also given for

Studies by earlier workers have shown differ-
ences in absorbance signals of platinum from
different platinum complexes when analysed
by flame AAS technique. For example, the
platinum absorbance of (NH4)t2PtCl4 (50 ppm)
is about twofold greater than that of K<2PtCl6
(Macquet and Theophanides 1974) and the
cis-DDP signal is about 12% higher than that of
H2PtCl6 (Macquet et al 1974). These differ-
ences are said to be due to the differences in
thermal stability of the complexes. The ability
of the more stable complex to survive as volatile
species in the flame is greater than the less
stable complexes; a volatile species is more easily
decomposed to free atoms hence a higher AAS
signal is obtained. (Macquet and Theophanides
the depression of platinum signals from H,PtCl(.
in the flame by several anions and cations,
H2PtCl6 is a compound which is formed
when Pt metal is dissolved in aqua regia and
diluted in hydrochloric acid after the evapora-
tion of nitric acid. The solution is used as
standard for platinum analysis using AAS tech-
niques because it is available commercially or
can be easily prepared.
In our preliminary attempt to analyse Pt
in serum and urine from patients who were
being treated with cis-DDP, using CFAAS, it was
found that the recovery of analysis was always
high (129-165%) (Haron, M.J. 1982). The
analysis and the recovery was done using
H2PtCl6 as standard. It was also found that at a
 MD. JELAS HARON AND WAN MD. ZIN WAN YUNUS
concentration of 500 ppb Pt in 0.05M HC1, the
absorbance peak height signal of cis-DDP was
higher by 38% compared to H2PtCl6.
This article describes the differences in
mechanism of the thermal breakdown between
H2PtCl6 and ciss-DDP solutions that possibly
occur in a carbon furnace atomic absorption
during the atomization process. This work was
carried out as part of the effort to optimize the
CFAAS technique for the analysis of platinum
in biological fluid i.e. serum, urine and cere-
bral fluid from patients after treatment with the
drug cis-DDP for various types of cancer
(Rozencweig et at 1977).
MATERIALS AND METHODS
Stock standard containing 1000 ppm Pt as
H,PtCl(i in 1M HC1 was obtained from BDH
Chemicals Ltd. Cis-DDP salt was obtained from
Mead-Johnson Ltd and an appropriate amount
was dissolved in 1M HC1 to give the 1000 ppm
Pt solution used as stock standard.
The curve for thermal breakdown of the
compounds from the solutions were recorded
using a Stanton Redcroft TG 750 thermobal-
ance which consists of a microfurnace and elec-
tronic balance, and a two-pen Speedomax strip-
chart recorder.
A platinum crucible (40 ml), which was
supplied by the manufacturers, was tared on
the microbalance at an argon flow rate of 20 ml
min'1. A sample solution (30 |ll) containing
about 3 mg of platinum as HtJPtCl(. or cis-DDP
was transferred into the crucible using an Eppen-
dorf micropipette. The solutions were prepared
by evaporation of 100 ml solution of 1000 ppm
platinum in 1M HC1 to 1 ml; the evaporation
was accomplished at a temperature between 90-
100°C. The thermogravimetric curve was re-
corded as the sample was heated at a heating
rate of 30°C min 1 from ambient temperature
to the maximum available tempreature of
1000°C.
A Philips x-ray generator equipped with a
Debye-Scherrer camera was used to record the
x-ray diffraction patterns of the residues formed
by heating the H.,PtClti solutions to about 350°C
in the thermobalance or in carbon furnace using
HGA 500 Perkin Elmer graphite atomiser.
The preparation of samples in the carbon
furnace for the x-ray analysis was done as fol-
80 PERTANIKA VOL. 12 NO. 1, 1989
lows: A stock solution of H2PtCl6 (lOjil) was
injected into the furnace and the solution was
dried at 100°C for 20 seconds with 20 seconds
ramp time. The temperature was increased to
350°C for 40 seconds with 40 seconds ramp
time. After cooling, the procedure was repeated
three times to obtain enough sample. The
residue was scraped from the furnace using a
stainless steel spatula and analysed by x-ray
method.
RESULTS AND DISCUSSION
Figure 1 shows a trace of the thermal breakdown
curve of H2PtCl6 as recorded by the thermobal-
ance. The pattern was similar to previously
reported data (Rowston and Ottaway 1979).
Schweizer and Kerr (1978) found that the
thermal breakdown of R;PtCl6. H,O salt to form
platinum metal occurred through several steps
as follows;
150°-180°C
H 2 PtCl 6 PtCl. + 2HC1
300°C-320°C
PtCL PtCl., + Cl,
375°C-510°C
PtCl2 Pt + CL
Plateaus A, B, C and D (Figure 1) corres-
pond to H2PtCl(., PtCl4, PtCl2 and platinum metal
respectively. In order to confirm the existence
of PtCI2, the residue from a thermogram on
heating the solution to the plateau C and from
carbon furnace on heating the solution to 350°
was analysed by x-ray powder diffraction, fol-
lowing the method published by Rowston and
Ottaway (1979). The d spacings of the x-ray
pattern were calculated and compared with the
standard values for PtCL, (The Powder Diffrac-
tion File 1972). The results showed that the
residue from the thermobalance at plateau C
and the carbon furnace contained only PtCL,.
A typical example of the x-ray pattern of the
residue from carbon furnace is shown in Table
L
PtCl2 is known to have different crystal-
line forms, a and p (Wiese et al 1970). The
latter, which is formed in this process (Schweizer
and Kerr 1978), contains discrete PtfiCl19 units
and is volatile (Schafer et al 1967); Landsberg
and Schaller, 1971). From the thermogravi-
metric traces, the number of moles of platinum
 DIFFERENCES IN THERMAL BREAKDOWN OF H,PtCl6 8c CIS-Pt(NH,)8Cl2 IN CFAA

metal and PtCl2 were calculated. It was found 7.00.

\
that 12% (mean) of the platinum is lost during \
the breakdown of PtCl 2 to Pt metal
(Haron 1982). However the amount of plati- 6.00,
\

\
num lost during the conversion varies with V /
sample size, heating rate and flow rate of the
inert gas used (Schweizer and Kerr 1978). \ AJ
5.00"
\ 1
TABLE 1
X-ray powder diffraction analysis of residue B
/
obtained after heating 30 ul aqueous 4.00-
solution of H^PtCl^ (0.3 mg Pt) in carbon /
furnace AAS to 350°C. X-ray conditions : J
source Cu K (1.54 nm) radiation 3.00 -
/ \ \
Lattice parameters (nm)
D
Calculated Literature* (PtOL,) 2.00 -
J \

0.673 0.669 \
0.400 0.403 1.00 -
0.374 0.378 700 600 500 400 300 200 !00
0.323 0.325
Temperature °C
0.292 0.293 Fig. I. Thermogravimetric curve (*) for a solution of
0.265 0.268 H^PtClh in hydrochloric acid. The residue on the
0.227 0.228 plete.au were HJHQt (A), PtClJB), PtCljC) and ft
0.199 0.200 metal (D). The line (*) is a temperature calibration.
0.187 0.187
0.176
0.164
0.177
0.163 7.00- 1
* The Powder Diffraction File, (1972).
\
*
I
On the other hand, the trace of ther-
6.00- J
mal breakdown curve of cis-DDP solution
shows that the compound is converted to
,oo- cis-Pt(NH3)
metallic platinum in the temperature range 2C12 /

290-500°C in a single step {Figure 2). This is F \

(i

similar to the thermogravimetric data on

\ J
res idue

\
4.00.
Pt(NH s ) 4 Cl 2 reported earlier by Kerr and
Chester (1971). They have shown that DDP,
which was formed as an intermediate during
the thermal breakdown of Pt(NH 3 ) 4 Cl 2 , was
I 3.00-
Pt me

i
converted to platinum metal in a single step,
as shown below:
2.00 "
105°C
Pt(NH3)4Cl2 »Pt(NH 3 ) 2 Cl 2 + 2NH 3 \

290°-500°C
5Pt(NH 3 ) 2 Q 2 > 5Pt + 4HC1 + 2N2 700 600 500 400 300 200 100
+ H2 + 6NH4C1 Temperature °C
From the thermogravimetric traces {Figure!) Fig. 2. Thermogravimetric curve (A) for ds-Pt(NHy)2Cl2 in a
it can be seen that cis-DDP was converted to hydrochloric acid solution. The line (•) is a temperature
metallic platinum without any significant loss. calibration.

PERTANIKA VOL. 12 NO. 1, 1989 81

 MD. JELAS HARON AND WAN MD. ZIN WAN YUNUS
CONCLUSION
In a carbon furnace, cis-DDP could also be con-
verted to metallic platinum in a single step,
whereas H9PtCl(. was shown to be converted to
PtCl2 before being converted to metallic plati-
num. The PtCl2 formed, which is volatile, could
be swept out by the flow of the argon gas (used
to protect the carbon furnace) in the same
manner as in the thermobalance. This could be
the explanation for the lower platinum signals
given by H2PtCl6 compared to those by cis-DDP
in CFAAS. It is therefore suggested here that
cis-DDP be used as a standard for analysis of
biological samples containing cis-DDP.
ACKNOWLEDGEMENT
The authors wish to thank Dr G.S. Fell, Royal
Infirmary, Glasgow and the late Dr. W.B.
Rowston,Glasgow College of Technology for
their assistance and for the provision of equip-
ment for the study.
REFERENCES
HARON, MJ. 1982. Analytical Investigation of Plati-
num and Zinc in Human Medicine Using Atomic
Absorption Spec trome try, Ph.D. Thesis,
Strathclyde University.
KERR, G.T. and A.W. CHESTER. 1971. Quantitative
Thermoanalysis of Evolved Ammonia. Application
to Ammonium Zeolite Y and Some Transition
Metal Ammine Chlorides. Thermochimica Ada 3:
113-124.
LANDSBKRG, A. and J.L. SCHALLER. 1971. The Ki-
netics and Equilibria of the Platinum-chlorine
System./ IJ>SS Common Metals 23: 195-202.
82 PERTANIKA VOL. 12 NO. 1, 1989
MACQUET, J.P. et ai 1974. Dosage du platine par
spectrometrie d'absorption atomique. Partie 1.
Mise au point d'une methode d'analyse dans les
complexes organometalliques et de coordination.
Analytica Chimica Ada 72: 251-259.
MACQUET, J.P. and T. THEOPHANIDES. 1974. Do-:
sage du platine par spectrometrie d'absorption
atomique. Part II. Resultats analytiques sur les
complexes organometalliques d'interet biolo-
gique: Effet cis-trans. Analytica Chimica Ada 72:
261-267.
MACQUET, J.P. and T. THEOPHANIDES. 1975. Ato-
mic Absorption and Relation between Stability
and Cis-trans Isomerism in Platinum Complexes.
Atomic Absorption Neivsletters 14(1): 23-25.
ROWSTON W.B. and J.M. OTTAWAY. 1979. Deter-
mination of Noble Metals by Carbon Furnace
Atomic Absorption Spectrometry. Analyst 104: 645-
659.
ROZENCWEIG, M. et ah 1977. Cis-diamminedichlo-
roplatinum (II)- A New Anticancer Drug. Annals
of International Medicine 86: 803-812.
SCHAFER, H, et al 1967. The Molecules PdBClI2
and PtgCljj in the Gaseous State. A Contribution
to the Problem of M6X12 Groups. Angeiu. Chem.
Internal Edit 6(3): 253-254.
SCHWEIZER, A.E. and G.T. KERR. 1978. Thermal
Decomposition of Hexachloroplatinic Acid. Inor-
ganic Chemistry 17(8): 2326-2327.
THE POWDER DIFFRACTION FILE. 1972. Joint Com-
mittee on Powder Diffraction Standards, Pennsyl-
vania, 19: 902.
WlESE, U. et ai 1970. X-Platinum (II) chloride
and Platinum (III) Chloride. Angew. Chem. Inter-
nal. Edit 9(2): 158.
(Received 31 May, 1988)
Pertanika 12(1), 83-88 (1989)

On the Higher Order Edge-Connectivity of

Complete Multipartite Graphs •

Y.H. PENG*, C.C. CHEN and K.M. KOH

Department of Mathematics
National University of Singapore
Kent Ridge, Singapore 05-11

Key words: kth order edge-connectivity; efficient k-separation; complete multipartite graph;
edge-toughness.

ABSTRAK
Biarkan Gsebagai suatu graf terhubung yangmempunyaip> 2 titik. Untuk k = 1, 2, ..., p - 1, kehubungan-
garisperingkat k yangdiberi lambangX(K) (G), ditakrifkan sebagai bilangan terkecil garis-garis yangapabila
dikeluarkan daripada G akan meninggalkan suatu graf yang terdiri daripada k + 1 komponen. Dalam artikel
ini kita akan menentukan kuantiti X(K) (GJ bagi sebarang graf muUipartit lengkap Gr Sebagai akibatnya
kita perolehi syarat perlu dan cukup supaya graf Gn dapat difaktorkan menjadi pohon-pohon janaan.

ABSTRACT
Lei G be a connected graph with p> 2 vertices. For k = 7, 2,..., p - 1, the kh order edge-connectivity of G,
denoted by X(K> (G), is defined to be the smallest number of edges whose removal from G leaves a graph with
k + 1 connected components. In this note we determine X<K) (GJ for any complete multipartite graph Gr As
a consequence, we give a necessary and sufficient condition for the graph Gn to be factored into spanning trees.

1. INTRODUCTION It is easy to see that for any tree T, XiK)(l^)

Let G be a connected simple graph of order p
and size q. Denote by V(G) and E(G) the vertex
set and edge set of G respectively. The edge-
connectivity X = X(G) of G is defined to be the
smallest number of edges whose removal from
G results in a disconnected or trivial graph.
This notion has a natural generalization. Fol-
lowing Goldsmith et al (1980), for each k = 0,
1» -.., p - 1, the kth order edge-connectivity of G,
denoted by X{K) (G), is defined as the minimum
number of edges of G whose removal increases
the number of components of G by k. Note that
X{{))(0) = 0, X^(G) = X(G) a n d X<>- ])(G) = q. T h e
properties of XiK)(G) were studied previously in
Boesch and Chen (1978), Goldsmith (1980 and
1981), Goldsmith et al. (1980) and Sampath-
kumar (1984).
= k. Furthermore, since any connected graph G
contains a spanning tree, XiK)(G) > k. It was
proved in Peng etaL (1988) that XiK)(Kf)=\ k(2r
- k - 1) for each k = 0, 1, ..., p - 1. In this note
we shall determine the kth order edge-connec-
tivity of a complete n-partite graph and then
use the result to derive a necessary and suffi-
cient condition for a complete n-partite graph
to be factored into spanning trees.
Throughout this article, we write Gn =
Ku(mv m,}, ..., mj, n > 2, to denote a complete
n-partite graph with n partite sets Vv K>, -., Vn
such that I V.\ = m. > 1 for each i = 1, 2, ..., n.
For the sake of convenience, we always assume
ml < ml} < ... < mn.
A graph G is called a complete multipartite graph
if G = Gn for some integer n > 2.

* On leave from the University of Agriculture, Malaysia.

 Y.H. PENG, C.C CHEN AND K.M. KOH

For those graph-theoretic terms used but two non-trivial components Q and Q,. Except
not defined here we refer to Behzad etal. (1979). for the two cases of separation shown in Figures
l(b) and (c) for n = 3 and rc = 4 respectively, it
2. EFFICIENT SEPARATION can be checked that there is always a partite set
Let G be a connected graph of order p, and k Vr of Gn such that (Q, u Q,) - V still consists of
be an integer such that 1 < k < p-\. Following two non-trivial components Q= Q- Kand Q^=
Goldsmith et al (1980) again, by an efficient k- (^ - Vrwhere Vf is separated into two sets V7and
separation of G, we mean a removal of \{K)(G) Vr in that separation. (Figure l(d)) Note thatVr or
edges from G so that G is separated into k + 1 Vr may be empty.
components. Call a component of a graph triu-
ial'xf it is a singleton, and non-trivial otherwise.
It was pointed out in Peng et al (1988)
that every efficient k-sepamtion of K(\ <k< p-
1) always results in at least k trivial components.
In this section we shall study the possible situa-
tions after performing an efficient separation
on Gn.
Let A and B be two subsets of V(G). We
denote by Ec (A,B) the set of edges of G each
W
joining a vertex of A to a vertex of B, and by
Figure 1.
eG(A,B) the number of edges in E(i(A,B). In
particular, we write e(,(A) for e(;(A,A)9 and e(i(v,B) So, the complete (n-l)-partite graph G =
for ^;({\)},B) where x> € V(G). The minimum Gn - Vf is separated into two non-trivial
degree of G is denoted by §(G), i.e. b(G) = components Q^and Q*. Let e denote the num-
min{degf.(u) I \) e V(G)). ber of edges removed in this separation of G\
First of all, we have and e* denote the number of edges deleted to
LEMMA 1. The number of edges of the graph separate Gn into Q and Q,. Then
Gn needed to be removed to separate Gn into two non-
trivial components is greater than 8( GJ, except when =e +
G - /( 2 (2,2), in which case, the number is equal to By induction hypothesis,

Proof We proceed by induction on n. For

the case n = 2, let G2, G2 * ^ ( 2 , 2 ) , be separated But
into two non-trivial components, and let e* m i f r
8(G') = ^n)- r * «
denote the number of edges removed in this 5(Gn) •*•„., if r - n,
separation. We may assume that both partite and
sets Vj' and V2 of G2 are divided into two sets.
Let V] be divided into a and b vertices, r-.nd V%
be divided into c and d vertices. {Figure l(a)j if r * n
Then a,£,cand dare positive. Since G 2 * K2(2,2)y m if r = n.
not all of them are equal to 1. Thus e* = ad + Therefore e* > 8(GJ, as required.
bc> a + b. If ad + be = a + by then c = d = 1 since It remains to consider the two exceptional
ay b, cy d are positive integers. This implies m2 cases.
= c + d = 2. Since m2 > m, > 2, r^ = 2. But this Case (i). The separation of G3 as shown in
contradicts our assumption that G2 * /£,(2,2). Figure l(b).
Thus, we have e* > a + £> = 8(G2). Let the partite set Vk be divided into a and b
Now, suppose that the statement holds vertices, and let e* denote the number of edges
for any graph G nl (n > 3). We shall show that removed in this separation. Then
the statement is also true for any Gn. Assume
that Gn * *^(2,2), and let Gn be separated into e* - mb + mm. + ma
1
» j j

84 PERTANIKA VOL. 12 NO. 1, 1989

 ON THE HIGHER ORDER EDGE-CONNECTIVITY OF COMPLETE MULTIPARTITE GRAPHS

= m.(b + m.) + ma of (^, we also separate H* into three compo-

> m. + m. (since b + m > 2, a > 1) nents. But this contradicts the minimality of
5(GJ,' VK)(GJ. The result thus follows.' •
Remark. We note that the result (ii) in
as required.
Theorem 1 can occur only when Gn = ^(Wpin.,),
Case (ii). The separation of G4 as shown in
where w., miy > 2.
Figure l(c).
Let e* denote the number of edges deleted in
3. HIGHER ORDER EDGE-
this separation. Then
CONNECTIVITY
e* = mm. + m.? + m.m/ + In this section we shall apply Theorem 1 to
= m.(m
' v
j
+ wi + m.m{ + determine the klh order edge-connectivity of
> m. + ?/i. + any complete ??-partite graph.
We begin with the following result.
LEMMA 2. Let 7 D V(Gn) such that I 71 = /
The proof is now complete. •
> 1 and e(. (T) + e(.Jl]Gn - T) = MHGJ. Then
We are now ready to prove the following
(i) there exists (O € T such that deg( {(0) = 5(G n ),
main result of this section.
and
THEOREM 1. Let p be the order of the graph
(ii) if jf = T- {\)| and G = Gn - \), where \J 6
G , and k be any integer with 1 < k < p-\. If Gn is
7; then ec.(3T) + e(.{T\G' - T) = X (t ; !) (G).
separated into components by an efficient k-separa-
M>te. By the assumption of Lemma 2, we
tion, then
are% indeed, given an efficient /-separation of G,
either (i) at kast k of the components are trivial, which separates it into / + 1 components j\l
or (ii) k - 1 of the components are trivial, and (x € 7) and Gn - T. The subgraph Gn - 7* must
the other two are K>- be connected as X(t)(Gn) < A,(t+1)(Gn)!'
Proof Suppose there are two non-trivial / V ^ (i) We suppose the contrary. Then
components Q and Q, of Gn after the removal no element of 7is in V or in any other partite
of XlK)(Gn) edges in an efficient k-separation of set V of Gn such that I V|l = I V I Let t> € Fand
Gn. We shall show that the induced subgraph F* =' r- I\i}. Consider the graph G n - F*. Note
is that Gn - r*is a complete multipartite subgraph
We first note that H is a complete multi- of Gn. So Vn is one of its partite sets. Let u e
partite subgraph of Gn. If H± £,(2,2), then by V and w e V* where V* is also a partite set in
Lemma 1, the number of edges removed to the partition of Gn - T*. Since I Vn\ > I V*\, we
separate H into two components Q and Q, is have
greater than 5(//). But S(//) is equal to the
number of edges removed to separate H into e( (u, G - 71*) < ^.(ti, G - r ^ .
n ( n
Si 'n
a trivial component {x>\, and a component H - Therefore (Figure 2)
\) where v e V(H) such that deg^-o) = 5(11).
iwl G
Thus Gn can be separated into k+\ components ^ ( T * U (Ml) + ^ < t * y ' n ' ^
by removing less than X{K)(Gn) edges. This
contradicts the definition of Xw(Gn). Therefore
Q, = Q= K2 and H= K,(2,2).
}, Gn - ( T
Now, suppose that there is another non-
trivial component Q^ of Gn after the removal of
X(K)(Gn) edges in an efficient k-separation of Gn.
Then, by the argument above, we conclude that
H which contradicts the minimality of X<0(Gn).
i=(Q*V QJ<i and K2 = (Q, U (X)Gu are all
Thus (i) follows.
isomorphic with ^(2,2). Thus, Q - K> and the
(ii) Since I T'\ = t- 1, by the minimality
number of edges removed to separate H* = (Q,
u
On u ft) i n t o three components Q, Q, and of A/'-'VG'), we have
(^ is six. However, if we delete all the five edges
of H* which are incident with the two vertices

PERTAN1KAVOL. 12 NO. 1, 1989

 Y.H. PENG, CC. CHEN AND K.M. KOH

this separation, at least k of the components are

trivial. By Theorem 1, there are at most two
non-trivial components Q and Q in any efficient
/^-separation of Gn, and //= (Q, u Q,> = ^(2,2).
But the number of edges removed to separate
H into Qi anc* Q> *s e c l u a l to t n e number of
edges whose removal separates //into a trivial
component and a /(,(1,2) component. This
completes the proof of the first part.
Figure 2. By Lemma 2(i), there exists \), € S such
that degrt (\),) = 8(Gn). If k > 1, let us write
S} = S-)v\\ and //, = G,, - *r By Lemma 2(ii),
we have

Note that //} is also a complete multipartite

graph, and \ S{\ = k - I. Thus by Lemma 2(i),
there exists u, e S, such that deg;/ (u,) = 8(// t ).
If & > 2, by using the same argument as above,
we conclude that for i = 3, 4, ..., k, SA contains
u such that deg;/, ^u) = 6(// 1 ) where // = H
j - U. The proof is now complete. •
We are now in a position to establish the
main result of this note.
THEOREM 2. Let p be the order of the graph
Figure 3. Gn. Then fork = 1,2,...,/>- 1,

If the equality does not hold, then (Figure 3)

for some Hk t of Gn.
= eJ(D + et.(T\ C- T) Proof By Lemma 3, there exists S ^ ^(^ n )
such that I S\ = ft and fc(l><Gn) - ^ ( S ) + e(js,
Gn - S). Since S= j\),, U2, ..., vj is such that for
i = 1,2,...,*, d e g / / ^ (a).)"= 8 ( / / , ) , we have
which is impossible. •
/;fj (S) u /;f; (,s; Gn - S)
Let //k = {Ho, Hv ..., ffj be a family of
subgraphs of Gn defined as follows: //0 = Gn and
for i = 1,2,...,*, H. = f/_, - u for some u e V(H
,) such that deg^ (u) ='8(//,). Thus,
We shall now apply Theorem 1 and
Lemma 2 to prove the following result. A (Gj = ^;(S)
LEMMA 3. Let p be the order of the graph G ,
and k be any integer satisfying \<k<p-\. Then there
exists a set of vertices S of G such that \ S\= k and

V as required. D
Furthermore, S = (t),, \)(), ..., t>k| = V(Gn) -
4. SPANNING TREE FACTORIZATION
// k w a member of some Hv of G .
It is known that a complete graph K can be
To prove the first part, we show
factored into spanning trees (indeed spanning
that, for each k* 1,2,..., ^ - 1 , there is an efficient
paths) if and only if pis even (see for instance,
^-separation of Gn such that, after performing
86 PERTANIKA VOL. 12 NO. 1, 1989
 ON THE HIGHER ORDER EDGE-CONNECTIVITY OF COMPLETE MULTIPARTITE GRAPHS
Behzad et al (1979), p. 168). In the following
theorem we give a necessary and sufficient condi-
tion for the graph Gn to be factored into span-
ning trees.
THEOREM 3. The complete n-partite graph
K{mv mr ..., mj can be factored into spanning
trees if and only if
) To prove the above theorem we shall make
m m = m - \\
- >
for some positive integer k.
COROLLARY. The graph K2(m,n) is spanning
tree factorizable for the following integers m and n
(n> m):
(i) m = 1, and n > 1;
(ii) m = 1 ( m o d 2 ) , m > 1, and n = m + 1;
(iii) m > 2, and n = (m ~ I ) 2 ;
(iv) m > 4 , and n = (m - l)(m- 2 ) / 2 ;
(v) m-ab and n = (ab~ l)(b- 1) where a and
b are integers > 2.
COROLLARY 2. (i) The graph Kn(tn - 1,
m,m,..Mm) is spanning tree factorizable if and only
if{n- 1) MiO (mod 2).
(ii) The graph Kn(l9m9m,...ftn) is spanning
tree factorizable if and only if nm = 0 ( m o d 2).
(iii) The graph Kn(m,m,...,m) is spanning tree
factorizable if and only if m = 1 and n is even.
Denote by (O(G) the number of compo-
nents of G. A subset X of E(G) is called an edge-
cutset of G if (O(G - X) > 1. Following P e n g et al
(1988), the edge-toughness of G, d e n o t e d by x x (G),
is defined as
T,(G) - min i * * 1 X is an
1 Graphs. SIAM j. Appl. Math. 34: 657-665.
' U(G-X)-I'
edge-cutset of Gr
The above definition of %X{G) is, as a matter of
fact, motivated by the following result due to
Nash-Williams (1961) and Tutte (1961) inde-
pendently.
THEOREM A. A connected graph G has s edge-
disjoint spanning trees if and only if I X| > 5((0
( G - X) - 1) for each X^ E(G).
It follows from Theorem A that a con-
nected graph G has k edge-disjoint spanning
trees if and only if Xj(G) > k. Thus Theorem 3
is an immediate consequence of the following
result.
PERTANIKA VOL. 12 NO. 1, 1989
THEOREM 4. xx(Kn(mv mv ..., m))
mm
i-,-
use of the following result which was obtained
in Peng et al. (1988) as a corollary of a more
general theorem.
For each i = 1,2,..., I V(G) \ - 1, we write
A\(G) - ^<M)(G) - \{I)(G).
THEOREM B. Let G be a connected graph of order
p and size q. If the sequence (A.(G) I 1 < i < p-\)
is non-increasing, i.e. A.(G) > A.+](G) for each i =
l,2,...,p - 2, thenX^G) = q/(p\).
Proof of Theorem 4. By Theorem B, we only
need to show that the sequence (A.(Gn) I 1 <
i < p - 1) is non-increasing. By Theorem 2,
A.(Gn) = Xil)(Gn) - A,(M)(Gn) = 8 ( H , ) . Note that
for i'= 1,2,...,/A, H. = H.*x - u where deg/fi.,(U)
= d(H. j) and t>. is adjacent to every vertex of H.
v except those in the partite set (of the parti-
tion of H. j) that u belongs to. So, it is clear that
for i » 1,2,...,^-1, 8(//) <8(// M ). Therefore, the
sequence (A.(Gn) II\ < i < p - \) = (8(//) I 0
< i < p - 2) is non-increasing. •
REFERENCES
BEHZAD M., G. CHARTRAND, and L. LKSNIAK-FOSTER.
1979. Graphs andDigraphs. Belmont: Wadsworth.
BOKSCH F.T. and S. CHEN. 1978. A Generalization
of Line Connectivity and Optimally Invulnerable
GOLDSMITH D.L. 1980. On the Second Order
Edge-Connectivity of a Graph. In Proceedings of the
Eleventh Southeastern Conference on Combinatorics,
Graph Theory and Computing, Boca Raton. Con-
gressus Numerantium 29: 479-484.
. 1981. On the N-th Order Edge-
Connectivity of a Graph. In Proceedings of the Twelfth
Southeastern Conference on Combinatorics, Graph
Theory and Computing. Vol. 1 Boca Raton, Con-
gressus Numerantium 32: S75-382.
GOLDSMITH D.L., B.MANVEL and V.FABER. 1980.
Separation of Graphs into Three Components by
the Removal of Edges. J. Graph Theory 4: 213-218.
NASH-WILLIAMS C.STJ.A. 1961. Edge-Disjoint
Spanning Trees of Finite Graphs. J. London Maths.
Soc. 36: 445-450.
87
 Y.H. PENG, C.C CHEN AND K.M. KOH

PENG Y.H., C.C. CHEN and K.M. KOH. 1988. On TUTTEW.T. 1961. On the Problem of Decompos
Edge-Toughness of a Graph I. SouthEast Asian ing a Graph into n Connected Factors./ London
Mathematical Bulletin 12(2): 109-122. Maths. Soc. 36: 221-230.
SAMPATHKUMAR E. 1984. Connectivity of a Graph
- A Generalization,/ Comb. Inf. and Sys. Sc. 9(2):
71-78. (Received 12 July, 1988)

88 PERTAN1KA VOL. 12 NO. I, 1989

Pertanika 12(1), 89-98 (1989)

On Robust Alternatives to the Maximum Likelihood Estimators

of a Linear Functional Relationship
MOKHTAR B. ABDULLAH
Department of Statistics
Centre for Quantitative Studies,
Universiti Kebangsaan Malaysia.
43600 UKM Bangi, Malaysia

Key words: Outlier; linear functional relationship; Theil's estimators; Ll-norm estimator.

ABSTRAK
Keiias ini membincangkan beberapa calon penganggar kukuh bagi menganggar hubungan fungsian linear
ringkas (SlfR). Penganggar kebolehjadian maksimum klasik bagi SIJ*R boleh terjejas dengan kehadiran data
terpencil. Ini ialah kerana ianya merupakan penganggar yang berasaskan min. Beberapa penganggar yang
berasaskan median bagi SIIR diselidiki. Dan calon-calon yang dipertimbangkan itu penganggar jenis Theil
dan penganggar norma-Ll terubahsuai didapati paling kukuh (iaitu, tak peka kepada data terpencil).

ABSTRACT
This paper discusses a number of candidates for robust estimators of a simple linear functional relationship
(SLFR), The classical maximum likelihood estimators of the SLFR can be affected fry the presence of possible
outliers. This is due to the fact that they are mean-based estimators. Some median-based estimators, of the SIFR
are examined. Among those considered Theil-type estimators and the modified Ll-norm estimator are found
to be most robust (i.e insensitive to the outliers).

1. INTRODUCTION: MODEL, there exists a consistent ML estimate of the

ASSUMPTIONS AND MAXIMUM SLFR parameters. The main results concerning
LIKELIHOOD (ML) ESTIMATORS functional relationship were discussed by Lin-
The model considered in this paper is a simple dley (1947), Madansky (1959), Sprent (1969),
linear functional relationship (SLFR). It speci- Solari (1969), Moran (1971), Kendall and Stuart
fies that two mathematical variable £ and X] are (1973), and Anderson (1980).
linearly related but observed with mutually With the assumptions on the error terms
independent, normally distributed errors 8and in (LI), the log-likelihood function may be
£, respectively. That is we observe written as
x. = (g + 5, <S~N(0,cO
/j Y) L = constant n logo 2 - 1 / 2 G 2
e.~N(0,
where
rj = a + /}£ t , i = 1 , . . . , n
There are (n + 4) parameters to be esti-
mated, i.e, a, p, a 2 , T*, and fcp.T.,1^ (or
equivalently r| p ...,r| t i ). The presence of the
incidental parameters^., i = 1, ..., n leads to incon-
sistencies of the classical ML estimates unless
an additional assumption is made about the
variance ratio X = T 2 /a 2 . That is, if X is known
 MOKHTAR B. ABDULLAH

In the SLFR an outlier may be generated

(1.2) from one of three types of contaminations, i.e,
(a) a contamination in the x's where the con-
where taminated error term £f is drawn from
2
, >(fand

or
(b) a contamination in the /s where
Then the maximum likelihood estimates of /?,
h\ > T2and
a, and %£. are

or
2.S\
(c) simultaneous contaminations in both the
*'s and / s .
(1.4) (i)
The presence of the outlying observation
a = v -/3x (1.4) (ii) can cause the fitted line to be dictated by the
wild observation resulting in failure to pass
through the bulk of the data (xfy) i = l,...,n.
and
£,= The nonrobustness (i.e, sensitiveness to the bad
observation) is due to the fact that & and $
and the maximized log - likelihood function is given in (1.4) (i)-(ii) are based on the means of
the observed values. Hampel(1974) pointed out
that the mean (as an estimate of a location
/. = constant - n log —•' parameter) is nonrobust because its influence
A
function is unbounded. Using a similar (i.e
A
(1.4) (iii) influence function) approach Kelly (1984)
where r = y -a -fix demonstrates that the influence functions of
h=x+a* d and j$ are unbounded and, therefore, these
ML estimates are nonrobust. In contrast, it is
It should be noted from (1.3) that for X = +°° well known that in the location problem the
or equivalently O2 = 0 or £. = x model (1.1) median is more robust than the mean because
reduces to a regression of y on x while for X * its influence function is bounded. Therefore,
0 or T2 = 0 model (1.1) becomes a regression median-based estimators of the SLFR may
of x on y, provide good alternatives to the ML estimators.
In the next sections, we present several
2. OUTLIERS IN THE SLFR candidates for the estimators of the SLFR which
In the regression (of y on x) model, the pres- are based on the median of the observations.
ence of a possible outlier is always associated A simulation study is performed to examine the
with a contamination in the dependent variable performance of these candidates and their
y. A contamination that occurs in yr may be due improvement over the mean-based estimators.
to the error term yr being drawn from some
heavy-tailed distribution, e.g, e,~N\0,ht) ,hf 3. ESTIMATORS BY GROUPING
> T2. Since no distributional assumption is METHODS
made for the regressor x (i.e, x is fixed) then In the history of SLFR, grouping methods have
the occurence of a possible outlier in the been developed as an alternative to the classical
variable may only be caused by a recording or maximum likelihood estimation procedure. The
typing error. main advantage of the grouping methods over

90 PERTANIKA VOL. 12 NO. 1, 1989

 ON ROBUST ALTERNATIVES TO THE MAXIMUM LIKELIHOOD ESTIMATORS OF A LFR
the classical one is the simplicity of the calcula-
tions and the dropping of the assumption of
normality.
One of the well known grouping methods
suggested in the literature was that of Bartlett
(1949). The method consists of arranging the
^values according to the ordered rvalues,
forming three groups, omitting observations in
the middle group, and joining the two centre
of gravity (i.e, the means) of the remaining
groups by a straight line to get the desired
slope .estimator. In an obvious notation, the 3-
grouped-mean estimators of a and j5 proposed
by Bartlett (1949) are defined as
A y$~ y \ /Q i \ /:\
and
(3.2) (ii)
y =n
•
and the ^values are first ordered according to
the magnitude of the x's, and
n\ = k (3.2) (iii)
n\ = n-k + 1
where k is the number of observations in the
first and third groups.
It is not necessary to take an equal number
of observations in each group especially when
nis indivisible by three. However, for the special
case of equally spaced ^-values, Bartlett shows
PERTANIKA VOL. 12 NO. 1, 1989
that the optimal choice is k = n/3 as this will
minimize the sampling variance of j3 -
Bartlett also shows that the estimators are
consistent if the grouping is unaffected by the
errors 5, i.e, the ordering according to the *.
is identical to the ordering according to the xr
In other words, for
i= l,...,n
it is assumed that
Prob{|sJ> c/2}=0 (3.3)
so that the estimators in (3.1) (i)-(ii) are con-
sistent.
In the formulation of the Bartlett 3-group-
mean estimators in (3.1) it was assumed that
there is no outlier in the x- or in the y- values.
However, some difficulties may arise in
using the grouping method if there is a possible
outlier in the data set. The possible difficulties
are
(a) the presence of the outlying observation
may cause the ordering of the jc-values no
longer identical to the ordering of the de-
values, i.e, the assumption in (3.3) may no
longer hold.
(b) the fact that the mean was used as the centre
of gravity of each group may result in the
nonrobustness of the Bartlett estimators.
In practice, however, it is difficult to
determine whether problem (a) does or does
not arise since the true values £. are unlikely to
be known. Even if £. are known, it is difficult to
guarantee that this problem does not occur
when an outlier is present. For problem (b),
the most appropriate choice is to adopt the
median as the centre of gravity of each sub-
group since it is more robust than the mean.
However, the robustness of the grouped-
median estimators may still not be guaranteed
if the presence of the outlier changes the or-
dering of the x-values. In this situation, only a
proper allocation of the observations in the
first and third groups can avoid the estimates
from being influenced by the outlier. Unfortu-
nately, the most appropriate choice of the
allocation, k, in this situation is not known. The
choice of k = n/3 was suggested for the no-
outlier situation and may not always be appro-
priate for the case where an extreme outlier
may be present.
91
 MOKHTAR B. ABDULIAH

The median-versions of (3.1) (i) and Another Theil-type estimator was devised
(3.1 (ii) may now be written as by Siegel (1982) which is based on a repeated-
median method. This approach starts with the
(3.4) (i) pairwise slopes as in Theil's method, takes the
xa — x , medians in two stages; first, at each point and
and then across points. That is we find
& H = med {y -px},i= l , . . , n (8-4)(») (i) ft - median
where ft is given by (4.1) yielding (n-1) slopes.
fj = med { },} . n . < i< n
(ii) fts = median {ft} (4.5) (i)
>~ = med {y}> 1< i< n\
which is the median of a set consisting n slopes.
*i = med { * J , n\ < i< n
e, = med { x } , 1 < / < n] Finally,

n x and n 3 are given by (3.2) (iii) and again the Ss= median {y-ftx} (4.5) (ii)
y-values are arranged in accordance with the or
ordered values of the xys.
fi v = median /' „} • (4.5) (iii)

4. THEIL'S AND REPEATED-MEDIAN w here a. is computed from (4.3).

ESTIMATORS
Perhaps one of the simplest methods and yet 5. Ll-NORM ESTIMATOR
may produce estimates which are more robust Harvey (1977) pointed out that for a simple
than the classical estimates is the one proposed linear regression model, the LI-norm regres-
by Theil (1950). sion (of y on x) which consists of minimizing
This method begins with all possible pairs
of data points, calculates the slope obtained (5.1)
from each point, and finally determines the
median of these slopes. Assuming that the x\s is perhaps the most natural generalisation of
are all distinct, define the sample median in the location problem.
We now extend this well known method
to the problem of estimating the SLFR
This yields slope values. The estimator parameters.
For the SLFR model in (1.1) the LI-norm
of the slope is then
estimators of a and /? are given by & and j
P, = median (4.2) (i) which minimize

The estimator & is then given by

&= median {j>,-prxt}. (4-2) (ii)
Brown (1982) suggests that the solution
Alternatively, in d r (4.2) (ii) can be to (5.2) i.e the Ll-estimates of a and b can be
obtained from a procedure proposed by Maritz obtained as follows: perform Ll-norm regres-
(1979) which is similar to the one that leads to sions of y on x and x on y, from which their
(4.1). Maritz derives the estimate of the inter- respective estimates of a and 6 are obtained. By
/ • computing the residuals from each fitted line,
cept term by calculating the intercepts for the required estimates 6t and $ a r e determined,
pairs, i.e, \d depending on which line i.e y on x or x on y,
has the smaller sum of absolute deviation". That
Then the intercept estimator is given by is, by performing the Ll-norm regression of y
on x we obtain and ^3 which minimize Ry
ar = median { a} (4.4)
where
92 BERTA\IK.\ VOL. 12 NO. 1, 1989
 ON ROBUST ALTERNATIVES TO THE MAXIMUM LIKELIHOOD ESTIMATORS OF A LFR
(5.3) (i)
and similarly, performing the LI-norm regres-
sion of x on y yields $ v and /} v which minimize
R, where
The required estimates of [a, /?] are then
given by either [ $ , , pv ] or [ $ v ) j$ ] according
to the smaller of
[tf, /?] (5.4)
However, Brown (1982) rejects the esti-
mates obtained from the criterion (5.4) for the
starting values because they may over- or under-
estimate a and p, respectively. Over- or under-
estimation of the parameters occurs when the
estimates chosen by the criterion in (5.4) are
the ones that are influenced by the outlier.
5.1. Modified LI-norm Estimator
In this section we propose a modification of the
criterion in (5.4). The rationale behind this
modification is to establish a new Ll-norm
criterion that will avoid the possibility of under-
or over-estimation of the parameters.
Under- or over-estimation of the para-
meters was due to the inclusion of the outlier's
residual in Rx and #, respectively. This has
resulted in choosing incorrect estimates since
the outlier's residual associated with the poor
fitted line will always be smaller than the one
associated with the best (or robust) fitted line.
To avoid such an incorrect choice for the
estimates, we devise a procedure in which the
largest residual is to be excluded from the overall
sum of absolute residuals in each direction. By
omitting the largest residuals which are usually
associated with the possible outlier the compa-
rison between Rx and R is now based on the
corrected sum of absolute residuals associated
with uncontaminated data points. Let the cor-
rected sums of absolute residuals correspond-
ing to tf and Rx be R\ and R\, respectively,
where
-ax-p<
(5.5) (i)
PERTANIKA VOL. 12 NO. 1, 1989
max {jx ,-(y . - a , ) / /J,|} (5.5) (ii)
T h e n the r e q u i r e d estimates of [a, p\ are
then given by either [ # , , / ? J or [&K,$K]
according to the smaller of
[/C R\]\ (5.6)
Note that the unconnected sums of abso-
lute residuals associated with a poor fitted line
will always have a larger value than the sum of
absolute residuals corresponding to the best
fitted line. This is because the poor fitted line
does not pass through the bulk of the data
while the best one should pass through most of
the data points. Therefore, the correct choice
for the estimate comes from the fact that the
smaller value of the sum of absolute residuals
should always be associated with the best fitted
line.
However, a proper choice for the esti-
mates based on the two sums of absolute
residuals may not matter very much if no con-
taminated data point or only a mild outlying
observation is present in the data set. This is
because choosing either one of the two sums
of absolute residuals will still lead to reasonably
good estimates.
In situations where several outliers are
likely to be present (5.5) (i)-(ii) can be modi-
fied so that more than one largest absolute
residual can be omitted. That is, if there are r
possible outliers in the data set then (5.5) (i)
can be written as
(5.5) (iii)
where
U, = y. - a y ~P,x, i= 1,..., n
and
Uj = the m-th largest absolute y-residuals, \Uy
and r is the number of the largest absolute re-
siduals to be eliminated.
Similarly, (5.5) (ii) can now be expressed in the
form
(5.5) (iv)
93
 MOKHTAR B. ABDULLAH

where The codes for the various estimators are as

U, =xt-(yt~ax)/px follows:
SLFR : The ML of SLFR given in (1.4) (i)-
and (ii)
u\ = the m-th largest absolute xi-residuals, BDYX : The Bartlett's group median estima-
tors given in (3.4) (ii)
We shall now compare the performance BMN : The Bartlett's 3-group mean estima-
of the modified Ll-norm estimator in (3.6) with tors given in (3.1) (i)-(ii)
the original Ll-norm estimators in (3.4). OL1 : The original Ll-norm estimators de-
Example: (Contamination in the y *s) fined by criterion (5.4)
This example presents the analysis of a data set MLl : The modified Ll-norm estimators de-
of 11 observations generated from the SLFR fined by criterion (5.6)
model in (1-1). The true parameter values are THEIL : Theil's estimators given in (4.2) (i)-
(ii)
a = p = 1 and 8.,£~N(O,O.5*), * = 1,.-.-., 10 SIEGEL: Siegel's estimators given in (4.5) (i)-
2
and £ n ~A((U0 ). (ii)
The data set are given in Table l(a). From the results in Table 1 (b), the mean-
To compute the Ll-norm estimates we based estimators, i.e, SLFR and BMN were badly
use NAG-routine E02GAF (based on the algo- affected by the extreme outlier in the data set.
rithm of Barrodale and Roberts (1973)). The median-based estimators such as MLl,
We now summarize the results for the BDYX, THEIL and SIEGEL performed reasona-
estimates from the various methods in the bly well for both types of contamination.
following Table l(b). For OL1, its performance against the
extreme outlier suffers from the incorrect
A
TABLE l(b)
choices of the estimates Ot and p which over-
Estimates estimate a and /?, respectively.
In order to make better comparisons
Estimator A t among the various starting estimators, let &{ and
pr 1 = 1,..., 100 represent the hundred values
SLFR 2.530 2.641
BMN 2.479 1.615 of the estimates of the SLFR parameters, and
BDYX 0.909 1.114 let the 'efficiency' of the various estimators to
OL1 3.050 2.586 be measured by the 'empirical mean square
MLl 0.819 1.013 error' criterion which is defined as
THEIL 1.059 1.073
SIEGEL 0.935 1.107

TABLE

i 1 2 3 4 5 6 7

X. - 5.605 - 3.966 - 2.956 - 1.772 -1.215 - 0.442 1.199

-3.857 . -3.197 -3.112 -1.378 -0.112 1.906 1.833

i 8 9 10 11

X. 2.984 3.200 3.675 4.288

i

Yi 2.868 4.060 5.003 22.366

94 PERTANIKA VOL. 12 NO. 1, 1989

 ON ROBUST ALTERNATIVES TO THE MAXIMUM LIKELIHOOD ESTIMATORS OF A LFR
where L=100, 6 is the estimate of the para-
meter 6.
The hundred samples were generated
under the following sampling situations;
a= j3 = 1, n- 11,-5 < £ < 5, equally spaced
and with increment 1.
The uncontaminated errors 5~ N{ 0, cr2)
wherea = 0.5, and£ r N(0,r*) , T =0.5. For
each contaminated sample it is assumed that
the contamination occurs either in the x'sor in
the 315.
For a contamination in xfi the values of
the xr outlying observation and the correspond-
ing uncontaminated yr observation are gene-
rated by
xr = £r + Sr where <5,~ A/(o,/i*)
and
y,= and
Similarly, for a contamination in y^ the
values of the yr outlying observation and the
corresponding uncontaminated xrare obtained
from
x r = £ r + Sr where 8r~N{0,o2)
J, = 1 + £ , + £,,£,-A/(«,* 2 ). and h* > o\
In each contaminated sample, it is as-
sumed that only a single outlier is present and
this outlier is located at an r-th position and r
is either fixed or random. The selected value of
a fixed ris r= 11, i.e, the outlier is to be located
at the last data point. For the case where the
location of the outlier is selected at random r
can take any value between 1 and n. The select-
ed values of h are
h = 2.0, 6.0 and 10.0
In this simulation study, the NAG-library
subroutine G05DDF was used to generate the
normal variates 8 and £, respectively. All com-
puter programs were written in FORTRAN
and executed on the DEC-10 computer system
at the University of Dundee.
6. SIMULATION RESULTS AND
DISCUSSION
Tables 2(a)-(b) demonstrate the performances
of the various estimators for the case of con-
PERTANIKA VOL. 12 NO. 1, 1989
tamination in x, while Tables 3(a)-(b) for the
contamination in yr.
The performance of an estimator is judged
from its MSE value with respect to the true
values of a and j3, where a = fi = 1. Estimator
A is said to perform better (or be more efficient
or more robust) than estimator B if the MSE va-
lue associated with A is smaller than that of B.
The percentage efficiency between two
estimators is defined as the reciprocal of the
(the lowest MSE) 1/www . .
ratio x 100% is also com-
(the largest MSE)
puted. For each estimator these efficiencies as
well as their ranks (where the smallest rank
corresponds to the smallest value of the effi-
A,
ciency, and (so forth) for & and p are dis-
played in the second and the third rows, respec-
tively.
Inspection of the results in Tables 2-3
illustrates that in a situation where there is no
possible outlier (i.e, h = 0.5), the classical esti-
mators such as SLFR and BMN perform well.
In the presence of an extreme outlier (for
h > 6) Theil-type estimators (i.e, THEIL and
SIEGEL) and ML1 seems to be most robust
among those considered. Its MSE values are
relatively smaller (and the percentage efficien-
cies are higher) than those of the other estima-
tors in all situations.
The rank averages of the percentage
efficiencies for & and p also agrees with the
fact that the most robust estimators are SIEGEL
and THEIL and they are closely followed by
ML1.
In situations where less severe outliers are
present the performance of OL1 seems to be
comparable to that of ML1. However, OL1 tends
to perform badly when more extreme outliers
are likely to be present. This was due to the
incorrect choices of estimates made by the OL1-
criterion in (5.4) which either over- or under-
estimates the true parameters. This can be seen
from the poor performance of OL1 in Tables
3(a)-(b) compared to those of ML1 and Theil-
type estimators.
The high values of MSE (correspondingly
the small values of percentage efficiencies) for
SLFR and BMN simply confirm the fact that
these mean-based estimators are very sensitive
to the presence of an extreme outlier.
95
 MOKHTAR B. ABDULLAH

TABLE 2(a): a = p = 1, n - 11
x - contamination; r - 11, fixed.
No Outlier Outlier - xn RANK RANK
h 0.5 2.0 6.0 10.0 AVE. AVE

Est. & $ a f a t a $ a * + $

SLFR MSE .054 .004 .082 .013 .251 .060 .273 .133
% eff 100 100 95 69 34 20 32 9
Rank 6.5 6.5 4 1 1 1 2 1 3.4 2.4 3.9
BMN MSE .054 .004 .084 .010 .230 .040 .332 .071
% eff 100 100 93 90 37 30 26 17
Rank 6.5 6.5 3 5 2 2 1 2 3.1 3.9 3.5
BDYX MSE 065 .008 .078 .012 .105 .041 .099 .041
% eff 83 50 100 75 81 29 88 29
Rank 5 1.5 7 3 4 3 6 4 5.5 2.9 4.2
OL1 MSE .074 .008 .087 .013 .109 .023 .161 .067
% eff 73 50 89 69 78 52 54 18
Rank 2.5 1.5 1.5 2 3 4 3 3 2.5 2.6 2.6
ML1 MSE .083 .007 .080 .011 .097 .015 .100 .012
% eff 73 57 97 82 87 80 87 100
Rank 2.5 3 6 4 5 5 5 6.5 4.6 4.6 4.6
THEIL MSE .074 .005 .087 .009 .085 .013 .101 .015
% eff 73 80 89 100 100 92 86 80
Rank 2.5 5 1.5 7 7 6 4 5 4.0 5.8 4.9
SIEGEL MSE .076 .006 .081 .010 .092 .012 .087 .012
% eff 71 67 96 90 92 100 100 100
Rank 1 4 5 6 6 7 7 6.5 4.8 5.9 5.3

TABLE 2(b): a = P = 1, n » 11
x( -contamination; r random.

No Outlier Outlier - x RANK RANK

0.5 2.0 6.0 10.0 AVE. AVE.

Est. A A $ d t A P A t a+p
SLFR MSE .054 .004 .103 .008 .239 .046 .264 .123
% eff 100 100 81 100 35 24 32 8
Rank 7 6.5 1 5 1.5 1 2 1 2.9 3.4 3.1
BMN MSE .054 .004 .097 .008 .239 .038 .363 .068
% eff 100 100 86 100 35 29 23 15
Rank 6 6.5 2 5 1.5 2 1 2 3.1 3.9 3.5
BDYX MSE .074 .008 .087 .012 .101 .037 .084 .023
% eff 83 50 96 67 82 30 100 43
Rank 5 1 5.5 1.5 5 3 7 4 5.6 2.4 4.0
OL1 MSE .077 .007 .089 .012 .108 .017 .135 .051
% eff 70 57 94 67 77 65 62 19
Rank 2.5 2.5 3.5 1.5 3 4 3 3 3.0 2.8 2.9
ML1 MSE .080 .007 .084 .011 .083 .011 .086 .012
% eff 67 57 100 73 100 100 97 83
Rank 1 2.5 7 3 7 6.5 6 6 5.3 4.5 4.9
THEIL MSE .075 .005 .089 .008 .104 .015 .094 .014
% eff 72 80 94 100 81 73 89 71
Rank 4 5 3.5 5 4 5 4 5 3.9 5.0 4.5
SIEGEL MSE .077 .006 .087 .009 .097 .011 .089 .010
% eff 70 67 96 89 85 100 94 100
Rank 2.5 4 5.5 4 6 6.5 5 7 4.8 5.4 5.1

PERTANIKA VOL. 12 NO. 1, 1989

 ON ROBUST ALTERNATIVES TO THE MAXIMUM LIKELIHOOD ESTIMATORS OF A LFR

As far as c o m p u t a t i o n a l aspect is con- T h e s e robust estimates can be used as

cerned, the Theil type estimators (i.e, T H E I L
and SIEGEL) a n d t h e Bartlett-median estima-
tor, BDYX, are m u c h easier to c o m p u t e than
the o t h e r estimators. In t h e case of L l - n o r m
estimator (either O L 1 or ML1) its formulation
was originally based on a linear p r o g r a m m i n g
scheme (see W a g n e r 1959) which is compu-
tationally c u m b e r s o m e . However, this difficulty
has b e e n r e d u c e d considerably with t h e pres-
ence of an efficient algorithm p r o p o s e d by
Barrodale a n d Roberts (1973).
7. C O N C L U S I O N
T h e n u m e r i c a l evidence shows that the Theil-
type estimators (i.e, T H E I L a n d SIEGEL) a n d
ML1 are most robust a m o n g those candidates
presented in the study. T h e practical advantage
of the Theil-type estimators over ML1 is their
simple c o m p u t a t i o n s . On t h e o t h e r h a n d , ML1
estimates can be o b t a i n e d efficiently when such
an algorithm set o u t by Barrodale a n d Roberts
(1973) is available.
starting values in o t h e r robust estimination pro-
cedures of the SLFR such as t h e M-estimation
of H u b e r (1964).
ACKNOWLEDGEMENT
This investigation is a part of Mokhtar b.
Abdullah's Ph.D thesis at the University of
D u n d e e , Scotland. T h e research was s u p p o r t e d
by the G o v e r n m e n t of Malaysia.
REFERENCES
ABDULIAH M.B. 1987. Robust Estimation of a
Linear Functional Relationship. Unpublished
Ph.D thesis, Dept. of Mathematical Science,
University of Dundee.
ANDERSON, T.M. 1984. Estimating Linear Statis-
tical Relationships. Ann. Statist. 12: 1-45.
BARRODALE, I. and F.D.K. ROBKRTS. 1973. An Im-
proved Algorithm for Discrete LI Approximation.
SIAMJ. Numer. Anal. 10: 839-848.

TABLE 3(a): X* = j5 =Cl, n = 11

y - contamination; r = 11, fixed

No Outlier Outlier - yr RANK RANK

h 0.5 2.0 6 10.0 AVE. AVE.

Est. A P & A A $ & a+ $

SLFR MSE .048 .004 .101 .017 .368 2.058 2.538 108.5
% eff 100 100 82 47 24 .5 3 .01
Rank 6.5 7 3 1 2 1 1 1 3.1 2.5 2.8
BMN MSE .048 .005 .100 .011 .384 .059 .889 .125
% eff 100 80 83 73 23 18 10 7
Rank 6.5 6.5 4 4 1 2 3 3 3.6 3.6 3.6
BDYX MSE .061 .007 .083 .104 0.22 .092 0.26
% eff 78 57 100 100 86 50 97 34
Rank 5 3 7 6.5 5 4 5 4 5.5 4.4 4.9
OL1 MSE .069 .008 .105 .016 .151 .051 1.403 1.279
% eff 71 50 79 50 59 21 6 .7
Rank 3.5 1.5 2 2 3 3 2 2 2.6 2.1 2.4
ML1 MSE .069 .008 .108 .013 .092 .016 .089 .015
% eff 69 50 77 61 98 69 100 60
Rank 1.5 1.5 1 3 6 5 7 5 3.9 3.6 3.8
THEIL MSE .069 .005 .088 .009 .107 .012 0.98 .011
% eff 69 80 94 89 84 91 91 82
Rank 1.5 5.5 5 5 4 6 4 6 3.6 5.6 4.6
SIEGEL MSE .067 .006 .085 .008 .090 .011 .090 .009
% eff 71 67 97 100 100 100 99 100
Rank 3.5 4 6 6.5 7 7 6 7 5.6 6.1 5.9

PERTANIKA VOL. 12 NO. 1, 1989 97

 MOKHTAR B. ABDULLAH

TABLE 3(b): a = )3= 1, n = 11

yr - contamination; r random

No Outlier Outlier - yr RANK RANK

0.5 2.0 6.0 10.0 AVE. AVE.

Est. d $ * A 6 f A

SLFR MSE 055 .004 .087 .010 .351 .134 1.065 3.520
% eff 100 100 84 70 23 5 8 .2
Rank 6.5 7 1.5 4.5 1.5 1 2 1 2.9 3.8 3.4
BMN MSE 055 .005 .087 .010 .343 .040 .936 .113
% eff 100 80 84 70 23 17 10 7
Rank 6.5 6 1.5 4.5 1.5 3 3 4 3.1 4.4 3.8
BDYX MSE 067 .009 .079 .019 .159 .061 .438 .223
%eff 82 44 92 37 50 11 21 3
Rank 5 1 3.5 1 3 2 4 3 3.9 1.8 2.9
OL1 MSE 072 .008 .074 .012 .092 .019 1.595 .810
% eff 76 50 98 58 87 37 6 1
Rank 4 2.5 6 2 6 4 1 2 4.3 2.6 3.5
ML1 MSE 078 .008 .075 .011 .080 .012 .092 .011
% eff 70 50 97 63 100 58 100 73
Rank 1 2.5 5 3 7 5 7 5 5.0 3.9 4.5
THEIL MSE 075 .006 .079 .007 .100 .007 .107 .008
% eff 73 67 92 100 80 100 86 100
Rank 2 5 3.5 7 4 7 5 7 3.6 6.5 5.1
SIEGEL MSE 074 .007 .073 .008 .097 .009 .100 .009
% eff 74 57 100 87 82 78 92 89
Rank 3 4 7 6 5 6 6 6 5.3 5.5 5.4

BROWN, M.L. 1982. Robust Line Estimation with
Errors in Both Variables./ Amer. Statist. Assoc. Ill
71-79.
HAMPEL, F.R. 1968. The Influence Curve and Its
Role in Robust Estimation./ Amer. Statist. Assoc.
69: 383-393.
HARVEY, A.C. 1977. A Comparison of Preliminary
Estimators for Robust Regression./ Amer Statist.
Assoc. 72: 910-913.
HUBER, PJ. 1964. Robust Estimation of a Loca-
tion Parameter. Ann. Math. Statist. 35: 73-101.
KELLY, G. 1984. The Influence Function in the
Errors in Variables Problem. Ann. Statist. 12: 87-
100.
KENDALL, M.G. and A. STUART. 1973. The Advanced
Theory of Statistics. London: Griffin.
LINDLEY, D.V. 1947. Regression Lines and the
Linear Functional Relationship./ Roy. Statist. Soc
SuppL 9: 218-244.
MARITZ, J.S. 1979. On Theil's Method in Distri-
bution-Free Regression. Austral. J. Statist 21: 30-
35.
MADANSKY, A. 1959. The Fitting of Straight Lines
when Both Variables are Subject to Error./ Amer.
Statist. Assoc. 54: 173-205.
MORAN, P.A. 1971. Estimating Structural and
Functional Relationships./ Multivar. Anal :232-
255.
SiEGEL, A.F. 1982. Robust Regression Using Re-
peated Medians. Biometrika 69: 242-244.
SOLARi, M.E, 1969. The 'Maximum Likelihood
Solution' of the Problem of Estimating a Linear
Functional Relationship./^ Statist. Soc. 5 3 1 : 372-
375.
SPRENT, P. 1969. Models in Regression and Related
Topics. London: Methuen.
THEIL, H. 1950. A Rank Method of Linear and
Polynomial Regression Analysis. I. Proc, Kon. Ned.
Akad. Wetensh. A 53: 386-392.
WAGNER, H.M. 1959. Linear Programming Tech-
niques for Regression Analysis, / Amer. Statist.
Assoc. 54: 206-212.
(Received 10 June, 1988)

98 PERTANIKA VOL. 12 NO. 1, 1989

Pertanika 12(1), 99-106 (1989)

Satu Kajian tentang Getaran Terusik ke atas Selaput Segi Empat

Sama dengan Menggunakan Teori Usikan.

ZAINUL ABIDIN HASSAN and SALWA BT. ABU BAKAR

Jabatan Fizik
Fakulti Sains dan Pengajian Alam Sekitar
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

MOHD. LOTFY B. ALI SABRAN

Jabatan Matematik
Fakulti Sains dan Pengajian Alam Sekitar
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

ABSTRAK
Dengan menggunakan teori usikan, kajian ke atas getaran selaput segi empat sama dibuat. la menunjukkan
bahawa andainya taburan jisim selaput tersebut tidak sekata maka akan berlaku perubahan pada bentuk
getaran dan juga frequensi yang terhasiL Bentuk getaran dan frequensi yang terhasil tersebut dapat diperi-
halkan oleh teori usikan. Dengan menggunakan komputer bentuk perubahan getaran tersebut dapat dilihat
dengan tepat.

ABSTRACT
A study of perturbed vibration on a square membrane using standard perturbation theory is conducted. It is
shown that uneven distribution of mass on the membrane would result in a different shape of vibration. This,
in turn, resulted in a change in frequency. Both of these can be described using perturbation theory. Hence,
a computer is used to draw the shape of the vibration.

1. PENDAHULUAN saian secara hampir bagi sistem yang sukar

Getaran di atas selaput merupakan satu feno
menon fizik yang dipelajari oleh mana-mana
pelajar fizik pada peringkat universiti. la me-
rupakan satu penyelasaian kepada persamaan
gelombang dengan syarat sempadan yang ter-
tentu. Getaran di atas selaput merupakan peng-
unggulan kepada sistem yang sebenarnya, se-
perti getaran pada permukaan gendang, kom-
pang, rebana dan mana-mana alat muzik yang
sepertinya. Penyelesaian ini mengandaikan
bahawa ketumpatan permukaan bagi selaput
tersebut adalah sekata.
Teori usikan adalah satu kaedah peng-
hampiran yang terkenal dalam matematik dan
fizik. la digunakan untuk mendapat penyele-
untuk mendapat penyelesaian secara tepat.
Kalaulah sistem yang ingin dikaji tersebut mem-
punyai persamaan dengan model yang mana
penyelesaiannya diketahui dengan tepat, maka
kaedah teori usikan boleh digunakan. Kaedah
ini sangatlah berguna kerana teori usikan boleh
digunakan peringkat demi peringkat sehingga
kepada darjah ketepatan yang diperlukan.
2. GETARAN ATAS SELAPUT
Pergerakan selaput diperihalkan oleh persama-
an gelombang iaitu (Pain 1975)
d V
77 ...d)
 ZAINUL ABIDIN HASSAN, SALWA BT ABU BAKAR DAN MOHD. LOTFY B. ALI SABRAN
dengan — = c1 dan c ialah halaju perambatan
gelombang di atas permukaan tersebut. Den-
gan menggunakan kaedah pemisahan pem-
boleh-ubah, penyelesaian \\f dengan di mana \j/
adalah sifar pada bingkai segiempat sama yang
mempunyai panjang pinggir L ialah
y/ = Asinim— Isin in-?- \exp( itet)
V /J / V Li J
Ini adalah gelombang pegun dalam dua
dimensi, dengan A sebagai amplitud gelombang,
m dan n adalah sebarang integer yang mempu-
nyai hubungan seperti berikut :
rt + nr
dengan k sebagai nombor gelombang, semen-
tara (0 mempunyai kaitan dengan k seperti
berikut :
(0= kf
oleh sebab itu
...(3)
9) £
Dari persamaan (2), y boleh ditulis seperti
berikut :
y/ = <p(Xjy) T(t)
di mana <(>(x, y) adalah fungsi yang bergantung
pada ruang semata-mata sementara T(t) adalah
fungsi yang bergantung pada masa semata-mata.
Oleh itu persamaan (1) boleh ditulis seperti
berikut :
9
n ^y i y) Yy J)"*\
Persamaan (4) adalah persamaan eigen dengan
<|>(x, y) sebagai fungsi eigen, G)2 sebagai nilai
eigen dan ——V sebagai operator eigen. Secara
fiziknya proses matematik di atas samalah
dengan mengambil gambar keadaan selaput
tersebut pada satu ketika di mana amplitud
getaran pada ketika itu ialah maksimum. Se-
mentara <|)(x, y) ialah fungsi yang mempunyai
pertalian berikut:
= — sin —"jsinf anZ} ...(5)
LJ \ L)
100
dengan j- sebagai faktor penormalan, Rajah
(1), menunjukkan bentuk selaput tersebut
untuk m = 2, n = 2.
Garis nod berlaku apabila <|>(x, y) = 0 iaitu
pada x= — dan y = —dengan i dan j se-
bagai sebarang nombor asli. Rajah (lb) menun-
jukkan garisan nod untuk mod getaran m = 2,
...(2)
n = 2. Sementara rajah (lc) ialah kontornya.
3. TEORI USIKAN
3.1 Teori Usikan ke atas Kes Tidak Degenerat
Katakan operator yang bertindak pada sistem
yang tidak terusik diberi oleh persamaan
(Gasiorowic 1974)
A 0 o
di mana H o(j)n dan E n adalah operator fungsi
eigen dan nilai eigen yang unggul di mana
penyelesaian diketahui. Maka bagi sistem yang
terusik ia boleh ditulis seperti berikut :
„ = E,,<t>tl
A A A
dimana H - H,,+ H'
H' = operator usikan.
Dengan mengembangkan <j) „ di dalam siri
<t>"rn iaitu
Maka perubahan nilai eigen keperingkat
pertama yang terhasil disebabkan oleh usikan
IT diberi oleh persamaan
E
> " WnHl$HdT; =< n\H'\n> = H ...(7)
'
dan
=
En-Em
Ini adalah hasil piawai dari teori usikan ke
peringkat pertama bagi sistem yang tidak
degenerat.
3.2 Teori Usikan ke atas Kes Degenerat.
Untuk sistem yang degenerat pula, fungsi eigen
yang tidak terusik terdiri daripada gabungan
linear
fungsi-fungsi eigen yang degenerat ter-
sebut. Iaitu (Anderson 1971)
PERTANIKA VOL. 12 NO. 1, 1989
 GETARAN TERUSIK KE ATAS SELAPUT SEGI EMPAT SAMA DENGAN MENGGUNAKAN TEORI USIKAN

di mana TV adalah bilangan kedegeneratan. V\ Dari persamaan (9), dengan mengambil Vn2 =
adalah pekali yang memperihalkan nisbah ga- 1, maka kita dapati
bungan 0° di antara satu sama lain. Dengan
mengembangkan (|>n, fungsi eigen yang terusik ...(12a)
di dalam sebutan <p"ti maka kita dapati persamaan
berikut : Oleh itu vektor eigennya ialah
ax
...(126)
V
1
di mana bnm adalah pekali untuk sebutan </>„', di eigenvektor di atas tidak dinormalkan. Dua nilai
dalam pengembangan 0,,. Dari persamaan di memberi dua nilai eigenvektor.
atas, perhatikan apabila m = n
4. USIKAN JISIM KE ATAS SELAPUT
E'tiV - n, BERGETAR
" III

4.1 Menentukan Operator Usikan

Persamaan (8) adalah persamaan matrik.
Getaran di atas selaput pada satu-satu ketika
Untuk sistem berdegenerat gandadua, persa-
diwakilkan oleh persamaan (4), iaitu
maan (8) boleh ditulis di dalam bentuk
V
V Persamaan di atas adalah untuk taburan
di mana a = < . Persamaan di jisim ke atas permukaan selaput yang sekata.
Katakan p tidak lagi sekata, sebaliknya terdiri
atas boleh dipermudahkan kepada
dari dua sebutan iaitu
a
as-En r> p^> p + p(x,y)
.-(9)
di mana p disebutan pertama menunjukkan
Persamaan (9) hanya benar apabila deter- taburan jisim yang sekata dan sebutan kedua
minan menunjukkan perbezaan taburan jisim yang
bergantung pada kedudukan iaitu p'(x, y). Di-
«.. -E. «i andaikan bahawa p'(x, y)«p. Oleh itu sebutan
,v = 0
T
L
Yang mana ini bererti p p + p(x,>-)
yang mana dengan menggunakan pengemban-
gan binomial ia boleh ditulis dalam bentuk
1
(*>y)
Persamaan (10) memberi perubahan nilai p + p(xyy) " p
eigen keperingkat pertama. Sementara E\ mem-
punyai dua nilai, yaitu dengan mengabaikan sebutan yang lebih besar
dari peringkat kedua ke atas
2En **„.+ an

V (an ~an)* p + p(x,3?) •=•£'-

p |/ p
Oleh sebab itu operator eigen / / , , - > / /

PERTANIKAVOL. 12 NO. 1, 1989 101

 ZAINUL ABIDIN HASSAN, SALWA BT. ABU BAKAR DAN MOHD. LOTFY B. ALI SABRAN

-T V2
di mana H= H+H =
p + P(* >y) D J
p( x) sin 2 m —dx
pL o
) T
p
c
>y)
p P
•f i—dy ...(15)

Katakan p'(x, y) adalah jisim titik yang berada

pada titik (xo, yo) maka;
d i m a n a H0 =
dan H =
u-y() ...(16)

4.2 Usikan Jisim Titik ke atas Kes Tak De-

dengan 5(x - xo) dan 8(y - yo) sebagai fungsi
generaL
delta Dirac, dan (I sebagai jisim usikan.
Dengan menggunakan operator usikan H se- maka persamaan (15) menjadi
perti persamaan (13) dan memasukkannya ke 4 ji r' ax
dalam persamaan (7), maka perubahan nilai D= -\ 8( x-x,,)sin 2m—dx
eigen boleh didapati yaitu pL L
pL <>
ny
J >n-dy

Adakan sedikit perubahan tatatanda D= i •sin n——

= < wi, m\H\m, m>
tetapi pL2 = M, yaitu jisim keseluruhan selaput
P(x,y)
= < m, m m> tersebut
P
. 7txtl
tetapi // J m, n > = - a> * I wi, n > D= •sin 2
M L L
P(x,y) Oleh sebab itu perubahan frequensi ber-
Acolin = 0)1 mn < m, m m, m> ...(14) laku sepertimana yang diberi oleh persamaan
Persamaan (14) mengatakan bahawa per- (14) i.e.
ubahan frequensi akan berlaku andainya se-bu- 4u
x sin sin
p{ >y) M L L
tan < m, m my m > ^ 0. la adalah . Persamaan di atas mempunyai nilai
positif jika p(x, y) adalah positif (iaitu penam- maksimum bila jisim usikan tersebut diletak di
bahan jisim) dan adalah negatif jika p(x, y) mana nilai
adalah negatif (iaitu pengurangan jisim). , Ttxo . ty<,
2
Perhatikan sebutan sin m sin n =1
L L
D = < m, m m> Iaitu di titik antinod dan tidak akan berlaku
sebarang perubahan frekuensi andainya jisim
m> tersebut diletak pada garisan nod. Nilai per-
ubahan frequensi ialah
ir / fx Ttx It 7ty»
= -pJ Aft)^ = 2(00 mm /— sin m —^sin m
0 0
VM L L
Dari persamaan (5)
Dengan kata lain semakin tinggi mod ge-
4 f f ; 7
Ttx taran, semakin besarlah perubahan frekuensi
— d x dy yang berlaku, makin besar nisbah jisim usikan
D=—^J p( x,y")sin2m — si
Li o o dengan jisim keseluruhan, dan semakin besar
Andainya p( x,y) = p I x)p maka jugalah perubahan frekuensi.
Tetapi

102 PERTANIKA VOL. 12 NO. 1, 1989

 GETARAN TERUSIK KE ATAS SELAPUT SEGI EMPAT SAMA DENGAN MENGGUNAKAN TEORI USIKAN

W
A (t) nxo 7Vyo di mana d..= < n, i |//|72, j> untuk i, j = 1, 2.
—=2 V M
[JL, „ sin m sin m
Persamaan (10) menunjukkan bahawa
iaitu perbezaan pecahan bagi setiap mod ge- nilai eigen yang degenerat akan berpecah pada
taran hanyalah bergantung pada nisbah jisim dua nilai baru. Iaitu
usikan dengan jisim keseluruhan sahaja. E = E\ + E,! dan E = Elt + E ]
Sementara perubahan bentuk untuk Dengan menggunakan FT sebagaimana
fungsi eigen diberi oleh persamaan persamaan (13), makaai; dapat dinilaikan, iaitu
\m, m> = Im, wi>° + Ya . In, ft > ... (17)
| • | ^ ^ j »H .mm \
(K x>y)
a = < n, i n, j >
mm p
dengan mengadakan perubahan tanda
< H in
dimana a ,,k . mm
(lihatper-
a"k =< nk, i
(K*>y)~H
E\ .-E -
samaan 8).
Oleh sebab itu
( v.y) a* =_!!<„*, ,-|
a
«.mm ft m, m>
P
0): Kx
n, ft|- m>
Untuk jisim titik seperti persamaan (16)
sin zk—dx dy ...(18a)
L
•(o: < «, k\ t, m>
pi:
«;; = «,: = — ^ J J p^yym n—
TtX,, .
col, n sin
Ky KX
K
y ( \
PL' sin k—^sin k—^sin n dx dy ...\ \Sb)
7tx(, L L L
sin in sin
L
Dengan memasukkan nilai «nkmm ke dalam
persamaan (17), maka siri untuk mode m,m L 2 p ....
yang terusik diperolehi. Didapati bentuk ge-
taran adalah berubah sepertimanayang terdapat sin 2£ sin !n——dxdy ...(18c)
dalam rajah (2a), di mana (la) adalah bentuk
getaran untuk mode (2, 2) tanpa usikan dan Oleh sebab itu persamaan (10) menjadi
(2a) adalah bentuk getaran untuk mod (2,2)
dengan usikan — =0.13 berada pada titik anti-
4 ob)
nod (0.25, 0.25) untuk L = 1. Sementara (2b) 2 V
ialah kontornya. untuk jisim titik p(x,v) x-x.)
4.3 Usikan Jisim Titik ke atas Kes Degenerat.
Perubahan kepada peringkat pertama pada nilai Maka persamaan (18) menjadi
eigen untuk kes degenerat diberi oleh persa-
= sin
maan (10) iaitu TT "w——sin k—
a
,, + a.,., M L L
E.= 2 sin n sin
—

± j — 5 — + • ; • • • ••• (10) sin ft -sin rc——

PERTANIKA VOL. 12 NO. 1, 1989 103

 ZAINUL ABIDIN HASSAN, SALWA BT. ABU BAKAR DAN MOHD. LOTFY B. ALI SABRAN

4 jicoi KX
KX * *).
) . > • v
laitu satu syarat di mana usikan tersebut tidak
-sin sin " n ... \i9c) menyebabkan simmetri sistem tersebut terturun.
M Sementara perubahan bentuk untuk
fungsi eigen ke peringkat sifar diberi oleh
sin n sin k persamaan
Perhatikan L
2
sin
k^-sm
L L
= a1 ...(20«)
Dengan mengambil Vmt - I, dan dari per-
samaan (12a), ia menjadi
maka —— = a ...(20ft)
a "2
yaitu a";; = a'\na™,
dan rf£ = ««<*£ f,! = -</{< -«*.*
maka persamaan (10b) menjadi
tetapi EH = 0 a"* = a a"2* dan a,, == CCHlta"*n
+1

a a K (a\m + 1)
+

E =
. ( « • „ •

GL
2 maka 0 = 0
2 „»*

(i ,,(«•„, •
± 1)

yaitu£ ( , = Oataupun E. = ati>{ a\m + l) ... (C2l) Dari (21) maka

Dengan kata lain, apabila diletakkan
-oc , a"
usikan jisim titik ke atas selaput, maka nilai
eigen akan berpecah pada dua nilai, sama ada
0 ( ;' = 0 ; ; t J + a , , f > ; , . . . ( 2 4 a )
£ „ = E\ ataupun £ „ = E°H + a^(cc~ + 0

Perhatikan E~, = a.}>{ a1,,,, + l)

Dan 0H |0,' > = 0 iaitu dua fungsi eigen yang
terhasil adalah berortogan di antara satu sama
lain.
dari (19b) dan (20a) Nisbah percampuran 0 „r, dan 0 r , di
dalam 0,' ditentukan oleh sebutan
KX,,
Act)';,,, = sin KXit .
M
sin /r sin
L I
+ sin 2k sin 2n | ...(22) a= KX . 7TV ,
sin n sin k
L L
disebabkan sebutan di sebelah kanan adalah Kx,, Ky,,
gandadua, oleh itu adalah sentiasa positif, maka a = 0 bilasin k-^- = 0 ataupun sin fr—— = 0.
El sentiasa positif. Apabila a sifar, maka
2
E „ sama dengan sifar hanya apabila
Kxt, /IX ffxfl
sin n—*• = sin « =0 atau sin n iaitu selaput tersebut bergetar dengan
7rv „ ?r>. ;rx mod 0 dan kedudukan jisim titik ialah pada
= sin n—— = 0 atau sin k = sin k =0 gansan nod untuk mod <p n r
L L L
Ky „ Ky „ Sementara untuk kes yang tiada peruba-
atau sin k—'•— = sin n—— = 0 han tenaga iaitu <j> „ ia adalah sama dengan
L L mode <(>", (lihat rajah 2).

104 PERTANIKA VOL. 12 NO. 1, 1989

 GETARAN TERUSIK KE ATAS SELAPUT SEGI EMPAT SAMA DENGAN MENGGUNAKAN TEORI USIKAN

fQ)
©
9 7TV TV

•CO" :
Rajah 2: 0 ,- —sin 2—sin ;?—
9 TTr TV -h
0" = _£_si n -£L£si n 2—
"' L L L

Rajah Ic. Kontor untuk getaran mod m = 2, n =2, tanpa usikan,

Rajah la. Hentuk getaran untuk mode m = 2, n = 2, tanpa

usikan.

Rajah 2a. Hentuk getaran untuk mod m = 2, n = 2 dengan

Rajah 1b. Hentuk garis nod untuk mode m = 2, n - 2. usikan * = 0.13 berada pada titik (0.25, 0.25)

PERTANIKA VOL. 12 NO. 1, 1989 105

 ZAINUL ABIDIN HASSAN, SALWA BT. ABU BAKAR DAN MOHD. LOTFYB. ALI SABRAN

frekuensi yang terhasil sentiasa lebih rendah

dari frekuensi bagi sistem tanpa usikan; tetapi
sebaliknyajika pengurangan jisim yang berlaku.

5. PENGHARGAAN

Pihak pengarang ingin merakamkan terima

kasih kepada ahli-ahli Jabatan Fizik UPM yang
memberikan cadangan-cadangan yang baik dan
teguran-teguran yang membina terutamanya
Prof. Madya Dr. Mohd. Yusuf Sulaiman.

RUJUKAN

FAIN, HJ. 1975. The Physics of Vibrations and Waves.

Rajah 2b. Kontor untuk getaran bagi mode m = 2, n = 2, dengan Great Britain: Unwin Brothers Ltd.
usiknn * = 0.13 berada pada titik (0.25, 0.25)
SRINIVASAMN P. 1982. Mechanical Vibrations Analy-
sis. New Delhi: Tata-McGraw-Hill Publishing
Company Ltd.

4. KESIMPULAN GASIOROWIC S. 1974. Quantum Physics. United

Kajian getaran ke atas selaput di atas menun- States of America : John-Wiley Sons.
jukkan bahawa andainya selaput tidak sekata, ANDERSON E.E. 1971. Modern Physics and Quantum
maka bentuk getaran yang terhasil akan be- Mechanics. Philadelphia: W.B. Saunders Company.
rubah. Ini juga merubahkan frekuensi sistem
tersebut Kalaulah jisim usikan itu merupakan
penambahan jisim pada mana-mana titik, maka
(Received 29 January, 1988)

106 PERTANIKA VOL. 12 NO. 1, 1989

Pertanika 12(1), 107-112 (1989)

COMMUNICATION IV
Lime Requirements of Highly Weathered Malaysian Soils

ABSTRAK
Beberapa kaedah penentuan keperluan kapur untuk tanah terluluhawa Malaysia diuji. Lapan siri tanah
dipilih, di mana lima daripadanya ialah Ultisol dan yang tiga lagi, ialah OksisoL Kaedah yang diuji ialah
pengeraman gml (dianggap sebagai kaedah rujukan), titration dengan Ca(OH)2 (3 kaedah) dan satu kaedah
berdasarkan Al bertukar ganti. Keputusan menunjukkan kaedah titration Dunn (1943) bertali rapat dengan
kaedah pengeraman. Penentuan keperluan kapur berdasarkan Al bertukar ganti memberi taksiran yang baik
untuk Ultisol, tetapi tidak untuk OksisoL Gerakbalas tanah terhadap GML mengikut waktu dan kadar
pengapuran juga dibincangkan.

ABSTRACT
Various methods of determining lime requirements for lueathered Malaysian soils were tested. Eight soil series
were selected, of which 5 were Ultisols and 3 were Oxisols. The methods tested were incubation with GML
(regarded as reference methods), titration with Ca(OH)2 (3 methods) and a method based on exchangeable
AL The result showed that the titration method of Dunn (1943) was well correlated with the incubation method.
Determination of lime requirement based on exchangeable Al gave a good estimate of lime requirements for
Ultisol, but not for OxisoL The responses of soils to GML over time and rate of application are also discussed.

INTRODUCTION which leaching and the continued use of acid

Most upland soils in Peninsular Malaysia are
extremely leached and weathered, and hence
dominated by clay minerals of the 1 : 1 type and
sesquioxides. These soils, of which the Ultisols
and Oxisols are the most widespread, are acidic
in nature, with pH values ranging from 4 to 5.
The soils are reportedly high in aluminium satu-
ration and base deficient, and the charge on
the exchange complex varies with the change
in pH (Tessens and Shamshuddin 1983). Lim-
ing with ground magnesium limestone (GML)
to raise soil pH to the level suitable for plant
growth is the standard practice to improve the
fertility status of the soils, except for rubber
and oil palm which are more tolerant to Al
toxicity.
It is often thought that under the condi-
tion of low pH, aluminium and maganese are
present in the soil solution in quantities phy-
siologically toxic to plant growth (Foy 1984).
There are many factors contributing to the in-
crease of soil acidity in Malaysian soils, among
forming fertilizers are often quoted. This par-
ticular aspect of soil chemistry has been dealt
with at length by Thomas and Hargrove (1984).
Many methods of lime requirement de-
termination have been proposed and tested.
These include those based on exchangeable Al
(Kamprath 1970,1984; Lathwell 1979), titra-
tion with a base (Dunn 1943; McLean et aL
1978; Gillman and Sumpter 1986) and incu-
bation with lime (Trans and van Lierop 1982).
But in recent years the method based on KC1
extractable Al (exchangeable Al) values has
been accepted and is popular for weathered
tropical soils. The relationship between the
amount of CaCOt^ required to neutralize a given
amount of exchangeable Al is given by the
equation CaCO3 equivalent (t/ha) = factor x
cmol (1/3 AP) kg1; the factor ranges from 1.5-
3.3 (Kamprath 1984), depending on soil che-
mical properties. Lathwell (1979) used a factor
of 2 to calculate lime requirements of soils in
South America. However, Gillman and Sumpter
 J. SHAMSHUDDIN
(1986) found that lime requirement determi-
nation based on exchangeable AF+ was suitable
for Ultisols containing substantial amounts of
AF+, but not for oxidic soils (Oxisoils).
The objectives of this paper were to com-
pare various methods of determining lime re-
quirements currently in use for weathered
tropical soils and to propose a quick and reli-
able method for determining lime requirement
of weathered Malaysian soils.
MATERIALS AND METHODS
Soils
Eight top soils (0-15 cm depth) of acidic and
highly weathered Malaysian soils were selected
for the study. These are the soils of Rengam,
Serdang, Bungor, Kuala Brang, Lanchang, Se-
gamat, Sg. Mas and Munchong Series, repre-
senting the most common sedentary soils in
Peninsular Malaysia, where rubber, oil palm and
food crops are mostly grown. The taxonomic
classification of these soils is given in Table 1.
The samples were air-dried, ground, sieved
through a 2 mm sieve and kept for treatment
and various analyses.
The soils under investigation have been
classified and studied in detail, especially with
respect to mineralogy and charge properties
(Tessens and Shamshuddin 1983). They are
either classified as Paleudult or Acrorthox,
except for the Munchong Series which is classi-
fied as Haplorthox (Table 1).
Analytical Procedures
Routine Analysis. Soil texture was determined
by the pipette method of Day (1965). Soil pH
TABLE 1
Taxonomic classification of the soils under study
SERIES PARENT MATERIAL
Rengam Granite
Serdang Sandstone
Bungor Shale
Kuala Brang Shale
Lanchang Granodiorite
Segamat Andesite
Sg. Mas Serpentinite
Munchong Shafe
108 PERTANIKA VOL. 12 NO. 1, 1989
(1:2.5) was determined in water and in 0.002 M
CaCl,2 after 1 day of equilibration. Cation ex-
change capacity (CEC) was determined by the
1M NH4 OAc; Na and K from the NH4OAc ex-
tract were determined by flame photometer,
while calcium and magnesium were determined
by atomic absorption spectrophotometer. Alu-
minium was extracted by 1M KC1 and deter-
mined colorimetrically and organic carbon was
determined by Walkley Black Method (Allison
1965).
Incubation. This method is a slight modi-
fication of the method proposed by Tran and
van Lierop (1982). Appropriate amounts of
ground magnesium limestone (GML) were
mixed with 100 g soils in perforated plastic
bottles to obtain a liming rate of 0, 2, 4, 6 and
8 tonnes/ha (on weight basis). The treatments
were watered twice weekly at the rate of 25 ml
with distilled water. The amount of water added
is equivalent to about 2000 mm rainfall per
year. Soil pH (in water and in 0.002M CaClJ
was determined at the end of 2 months. The
amount of GML needed to increase soil pH
(pH water) to 5.5 was then determined.
Two soils, namely the Segamat and
Rengam Series were selected for further inves-
tigation where the soils were limed at 2 tonnes/
ha (on weight basis) and pH (H90) was deter-
mined after 0, 1,2, 3, 4, 5, 6, 7 and 8 weeks.
Titration (Method A). This method was
proposed by McLean et ai (1978). Ten g soil
in 50 ml water was shaken for 1 hour. It was
then titrated with 0.08M Ca(OHJ to pH 7.2.
Each titration for different soils was done in
duplicate. Lime requirement was estimated from
the amounts of Ca(OH)2 used.
FAMILY
Clayey, kaolinitic, Typic Paleuduli
Loamy, kaolinitic, Typic Paleudult
Clayey, kaolinitic, Typic Paleudult
Clayey, kaolinitic, Typic Paleudult
Clayey, kaolinitic, Typic Paleudult
Clayey, oxidic, Typic Acrorthox
Clayey, oxidic, Typic Acrorthox
Clayey, kaolinitic, Tropeptic Haplorthox
 LIME REQUIREMENTS OF HIGHLY WEATHERED MAIAYSIAN SOUS
Titration (Method B). This method was
proposed by Dunn (1943). In this experiment,
which was also replicated twice, appropriate
amounts of 0.08M Ca(OH)2 were added to 10
g soil in a plastic vial, so that it had 0, 5, 10, 15,
20 and 25 ml of Ca(OH),, Water was added to
make 25 ml in volume of the solution. After 1
hour of shaking and 2 days of equilibration, pH
was determined. A graph of volume of Ca(OH),,
was plotted against pH, and lime requirement
to bring the pH to 5.5 was estimated from the
graph.
Titration (Method C). This experiment was
conducted following the method of Gillman
and Sumpter (1986). Four g soil was put in 100
ml centrifuge tube. Forty ml 0.1M CaCL, was
then added and the contents shaken for 2 hours.
The samples were centrifuged and de-canted
and two more washings with 0.002M CaCl2 were
carried out. The pH of the third suspension was
adjusted to 3.5 with 0.1M HC1. One ml 0.08M
Ca(OH),, was added and shaken and pH was
determined after 2 days. The addi-tion of 0.08M
Ca(OH)2 (same rate) was conti-nued at 2-day
intervals until pH 7 was reached, with the pH
being recorded each time. A graph of Ca(OH),,
was plotted against pH, and lime requirement
to bring the pH to 6.0 was determined there-
after.
Exchangeable A I. Lime requirement of the
soils was calculated by the formula :
L.R (tonnes/ha) = Al cmol (p+) kg1) x 2
following the proposal of Lathwell (1979).
RESULTS AND DISCUSSION
Chemical Properties
Table 2 summarizes the chemical properties of
the studied soils. It is seen that pH is low with
values less than 5 and so are the basic cations,
while Al saturation is very high in the Ultisols
with values greater than 70%. Of particular
interest is the exchangeable aluminium of soils
of Kuala Brang and Bungor Series, in which
the respective values are 5.31 and 3.14 cmol
(p+)kg' (Table 2). The high values are related
to the presence of some weathereable minerals
in the soils. X-ray diffraction investigation (not
shown) showed that both soils contain some 2:1
clay minerals (vermiculite).
PERTANIKA VOL. 12 NO. 1, 1989
Soil Buffering
At the GML rate of 2 tonnes/ha, it was noted
that pH was quickly raised to over 5.5 in a week
for the Ultisol (Rengam Series) and then
remained at 5.5-6.0 up to 8 weeks (Fig. 1). In
the Oxisol (Segamat Series), there was a lower
increase in pH although the pH at the onset of
the experiment was higher than the Ultisol.
The pH of the Segamat Series soil remained
constant from second to seventh weeks, after
which it began to decrease. The buffering action
of Rengam Series soil which is dominated by
kaolinite is probably controlled by Al (Sham-
shuddin and Tessens 1983), while that of Se-
gamat Series is probably controlled by sesqui-
oxides (Gillman and Sumpter 1986).
Fig. 1: Weekly pH changes of soils undergoing GML
treatment
Fig. 2 gives the change in soil pH after 8
weeks with different rates of GML application.
Two Ultisols (Lanchang and Kuala Brang Series)
and two Oxisols (Sg. Mas and Segamat Series)
were presented here to show their response to
GML treatment. It was earlier deduced (Fig. 1)
that soil with low exchangeable Al (Segamat
Series, Table 2) was buffered mainly by the
oxides and that the pH began to decrease after
7 weeks. In Fig. 2, pH of Segamat Series was
found to be higher than other soils at 2, 4 and
6 tonnes/ha GML application.
The soil of Kuala Brang Series was found
to be the most buffered soil. This is due to the
presence of exchangeable Al, which is more
109
 •

TABLE 2
Selected chemical properties of the soils under study

PH (1:2.5) Basic Cations (cmol(p+)kg-') Al CEC Texture Analysis (%)

c
Series
H20 CaCl2 Na K Mg Ca (cmoUp+Jkg1) O.M.(%:) (c mol(p+)kg') Clay Silt Sand
PER

Rengam 4.3 3.8 0.09 0.08 0.11 0.23 1.09 1.98 4.90 39.2 17.6 42.2
Serdang 4.6 3.8 0.06 0.05 0.06 0.02 0.86 1.05 3.25 26.3 2.7 71.0
2 Bungor 4.1 3.6 0.07 0.15 . 0.13 0.24 3.14 1.45 6.00 37.9 12.5 55.6
Kuala Brang 4.4 3.9 0.06 0.17 0.07 0.18 5.31 2.03 9.50 5 *i' 22.4 33.5 44.1
o Lanchang 4.5 4.1 0.06 0.15 0.18 0.12 1.36 2.78 8.93 ' 56.1 6.9 37.0
r Segamat 4.9 4.6 0.04 0.12 0.96 1.10 0.31 3.00 9.18 86.0 10.2 3.4 %
Sg. Mas 4.6 3.9 : 0.04 0.18 0.46 0.24 0.62 1.60 •1 i 7.18 45.2 17.9 36.9
5
JO. 1

Munchong 4.4 4.1 0.06 0.30 0.14 0.84 1.51 •:. 2.41 1 '•*.. 9.05 68.8 9.8 21.4
 LIME REQUIREMENTS OF HIGHLY WEATHERED MALAYSIAN SOILS
than 5 cmol(p + )kg' 1 . This soil needs more than
8 tonnes/ha GML to bring the pH to 5.5. The
soil of Lanchang Series, which containes less
exchageable A1(1.S6 cmol (p+)kg"}, was less
buffered than the Kuala Brang Series.
Fig. 2: The change in pH with different rate of GML
application
Lime Requirement
It is found that there exists a good correlation
between the incubation method and the titra-
tion method of Dunn (1943), as seen in Table
4. The correlation between incubation and other
methods of lime requirement is not significant,
including the method based on exchangeable
Al proposed by Lathwell (1979). The relation-
ship between the method of Gillman and
Sumpter (1986) and the method based on
exchangeable Al is also good (Table 4). The
relationship between the method of McLean et
al (1978) and the method of Dunn (1943) is
given by the equation:
Y= 1.82 + 0.82X
R = 0.84, P < 0.01
It is seen that the Lathwell (1979) method
is able to estimate the lime requirement of Ulti-
sols (Rengam, Serdang) accurately. However, it
can not estimate the lime requirement of Oxisols
(Segamat, Sg. Mas). This is similar to what has
been found by Gillman and Sumpter (1986)
for the highly weathered soils of Northern
Queensland, Australia. The methods of deter-
PERTANIKA VOL. 12 NO. 1, 1989
mining lime requirement based on exchange-
able Al as proposed by Kamprath (1970 and
1984) and Lathwell (1979) are only suitable for
the Ultisols, where the amounts of Al in the
soils are rather high. The Ultisols, with kaoli-
nitic mineralogy, contain more Al than the
Oxisols (Table 2). This paper suggests that the
method of determining lime requirement
proposed by Dunn (1943) can be adopted to
determined lime requirements of Oxisols and
Ultisols in Malaysia. For the Ultisols with high
amounts of Al, however, the method of Lath-
well (1979) can also be used.
CONCLUSION
In the Ultisol, pH increases to above 5.5 in
about a week and remains at this value for more
than 8 weeks. In contrast, for the Oxisols pH
started to go down after 7 weeks. The best
method for determining lime requirements of
weathered soils (Ultisols and Oxisols) is the
titration method of Dunn (1943). For Ultisol,
the method of Lathwell (1979) which is based
on exchangeable Al is acceptable.
ACKNOWLEDGEMENT
The author wishes to record his appreciation to
UPM for funding this research and to the
laboratory assistants in the Soil Mineralogy
Laboratory for their help during the period of
the research.
J.SHAMSHUDDIN
Soil Science Department,
Faculty of Agriculture,
Universiti Pertanian Malaysia,
43400 UPM Serdang, Selangor Darul Ehsan,
Malaysia.
REFERENCES
ALLISON, L.E. 1965. Organic carbon. In Methods
of Soil Analysis, ed. C.A. Black. Agronomy Mono-
graph 9 : 1376-1379.
DAY, P.R. 1965. Particle Fractionation and Particle
Size Analysis. In Methods of Soil Analysis, ed. C.A.
Black. Agronomy Monograph 9 : 545-567.
DUNN, L.E. 1943. Lime Requirement of Soil by
Means of Titration Curve. Soil Sci. 56 : 341-351.
FOY, C D . 1984. Physiological Effects of Hydro-
gen, Aluminium, and Manganese Toxicities in
111
 J. SHAMSHUDDIN

Acid Soils. In Soil Acidity and Liming, ed. F. Adams. Double Buffer and Quick-test Features. Soil Sci.
Agronomy Monograph 12 : 57-98. Soc. Am.j. 42: 311-316.
GILLMAN, G.P. and E.A. SIMPTER. 1986. Surface SHAMSHUDDIN, J. and E. TESSENS. 1983. Poten-
Charge Characteristics and Lime Requirements tiometric Titration of Acid Soils from Peninsular
of Soils Derived from Basaltic, Granitic and Malaysia. Pertanika 6 (1): 71-76.
Metamorphic Rocks in High-rainfall Tropical TESSENS, E. and J. SHAMSHUDDIN. 1983. Quanti-
Queensland. Aust. J. Soil Res. 24: 173-192. tative Relationship between Mineralogy and Properties
KAMPRATH, EJ. 1970. Exchangeable Aluminium of Tropical soils. Serdang: UPM Press.
as Criterion for Liming Highly Leached Mineral THOMAS, G.W. and W.L. HARGROVE. 1984. The
Soils. Soil ScL Soc Am. Proc. 34: 252-254. Chemistry of Soil Acidity. In Soil Acidity and Lim-
KAMPRATH, EJ. 1984. Crop Response to Lime in ing, ed. F. Adams. Agronomy Monograph 12: 3-
the Tropics. In Soil Acidity and Liming, ed. F. Adams. 56.
Agronomy Monograph 12: 349-368. TRANS, T.S. and VAN LIEROP, W. 1982. Lime Re-
LATHWELL, D.J. 1979. Crop Response to Liming quirement Determination for Attaining pH 5.5
of Ultisols and Oxisols. Cornell International Agri- and 6.0 of Coarse Textured Soils using Buffer-pH
cultural Bull. 35. Ithaca: Unversity of Cornell. Methods. Soil Sri. Soc. Am. J. 46: 1008-1014.
MCLEAN, ED., D.J. ECKERT, G.Y. REDDY and J.E.
TRIEKWELER. 1978. An Improved SMP Soil
Lime Requirement Method for Incorporating (Received 17 December, 1987)

<

112 PERTANIKA VOL. 12 NO. 1, 1989

 P e r t a n i k a l 2 ( l ) , 113-117 (1989)

COMMUNICATION V
Effects of Effective Size in Rapid Sand Filtration

ABSTRAK
Kajian mengenai proses penurasan pasir laju telah dijalankan di Loji Rawatan Air Sungai Langat. lanya
bertujuan untuk menilai kesan saiz efektif media pasir terhadap kekeruhan, masa operasi dan kadar atiran.
Kertas kerja ini mengutarakan penemuan daripada kajian tersebut. Hasil daripada kajian didapati tiada
perubahan yang ketara dalam nilai kekeruhan selepas dituras untuk saiz efektif di antara 0.4 - 0.9 mm.
Walau bagaimanapun masa operasi bertambah untuk saiz efektif yang lebih besar. Di samping itu, masa
operasi juga dipengaruhi oleh perubahan kadar aliran, di mana pertambahan kadar aliran sebanyak 50%
telah menghasilkan pengurangan masa operasi selama 2 jam untuk setiap saiz efektif yang dikaji.

ABSTRACT
A study on the effects of effective size of sand media in a rapid filtration process with respect to turbidity, filter
run and flow rate was carried out at Sungai Langat Water Treatment Works. The findings of the study show
no significant difference in the final turbidity achieved with effective sizes (E) range between 0.4 and 0.9 mm.
However, filter run increases with higher effective sizes. Filter run is also affected by a change in flow rate
a 50% increase in flow rate results in the reduction of filter run by two hours for the effective sizes studied.

INTRODUCTION a rapid filtration process with respect to tur-

Rapid filtration is mainly a physical and chemi-
cal process for removing suspended impurities
from water by passage through porous media
normally by action of gravity resulted from the
head difference. The main processes involved
can be summarised as straining, flocculation
and sedimentation.
Fitration rates of a rapid sand filter lie in
the range of 120-300 mVmVday (Fair et
al 1968 and Clark et al 1977). The higher
filtration rates enable the space requirement
for a rapid filtration unit to be reduced to as
much as 20% of that required by a slow sand
filter (Schulz & Okun 1984). This is, however,
achieved at the expense of the would be bio-
logical process of a slow sand filter.
Rapid sand filtration is not a new process
and in fact almost all the major water treatment
plants in Malaysia employ such a process unit.
However, the appropriateness of the filter media
and the efficiency of the process can still be up-
graded. The study undertaken aims at obtain-
ing the effects of effective size of sand media in
bidity, filter run and flow rate.
The study was conducted at Sungai Lang-
at Water Treatment Works utilizing the chemi-
cally treated settled water prior to the rapid
sand filtration.
MATERIALS AND METHODS
The filter column utilized in the study is fabri-
cated from clear perspex and equipped with
backwashing facility as shown in Figure 1. The
column is designed to operate as a gravity flow
process with regulating valves adjusted manu-
ally to compensate for the increasing head loss
and thus maintaining a constant rate of flow.
Initially, sands of various effective sizes
(0.4, 0.6, 0.8 and 0.9 mm) and uniformity
coefficient (U) of 1.6 were studied to obtain
optimum depths of filtration for each effective
size. Common range in U for single-medium
sand filters is 1.3 to 1.7 (Hammer 1986). The
optimum depth is defined here as the depth of
the sand at which turbidity begins to stabilize
i.e. no significant reduction takes place. The
 AHMAD JUSOH, MEGAT JOHARI MEGAT MOHD. NOOR AND ABDUL HALIM GHAZALI

various effective sizes can be obtained from the

particle size distribution curve as shown in Figure
2. Also shown in the figure is an example of
how to obtain a particular effective size and
uniformity coefficient, which in the example
are 0.4 and 1.6 respectively. The filter column
was operated at the flow rate of 7.1 mVmVhr
until the effluent turbidity becomes stabilized
which, from experience normally occurs after
30 minutes. The respective optimum depths of
filtration for the various effective sizes are to
be utilized in the second stage of the study.
Turbidity was the only parameter being mea-
sured at this stage.

0-5 0 7 I
Size of separation , {

Fig. 2: Particle size distribution of sand media

depths. For all affective sizes, turbidity reduces

gradually and later flattens out as it approaches
the optimum depth. The smaller the effective
size the lower is the effluent turbidity achieved.
However, at optimum depth the difference is
insignificant. It can be seen that optimum depth
reduces with the reduction in effective size as
shown by the line of optimum depth in the
figure.

Fig. 1: Layout of the rapid sand filtration system

The second stage involved operating the

filter at the optimum depths for all the effective
sizes mentioned earlier for a total running time
that is, until the measured filter head loss was
2.5 m, as per normal practice (Linsey 1979).
Backwashing was then carried out prior to the
subsequent operations. Two different flow rate
were used, that is 4.7 mVm 2 /hr and 7.1 mY
mr/hr. Turbidity, head loss and filter run were
monitored.
RESULTS AND DISCUSSION
Figure 3 shows the relationships between the
effective sizes and their respective optimum
30 40 50
Depth of Media ( cm)
Fig. 3 Relationship between effluent turbidity and depth of media
for various effective sizes.
Figures 4 to 11 show the variations of turbi-
dities and head losses with respect to filter run
at flow rates of 4.7 mVm 2 /hr and 7.1 mV'mV
hr. Effluent turbidity was not affected by the
fluctuation in influent turbidity of less than 5

114 PERTANIKA VOL. 12 NO. 1, 1989

 EFFECTS OF EFFECTIVE SIZE IN RAPID SAND FILTRATION

NTU which was the maximum turbidity value 300

recorded while conducting the study for all the

effective sizes. In fact the effluent turbidities lie 2 50

within the narrow range of 0.1 to 0.2 NTU -200

\ lnfb«nt furbWty

which is not significantly different. This is within 1

the limit recommended by WHO for drinking

water quality which is less than 1 NTU for l,oo Flow r at* = 4-7m"/m*/hr

effective disinfection (WHO 1984). The filter E = 0 6 mm

run increases with the increase in effective size. ^——___^_ «fflu«n» turbidity

Higher effective sizes will therefore allow fewer 0

Runrw>g T i m * I hrc )
backwashings and subsequently reduction in the
operation cost due to a longer filter run. As Fig. 6: Variation of influent turbidity, effluent turbidity and
expected, for all effective sizes the head loss in- head loss with respect to filter run
creases gradually at the earlier stages and rapidly
towards the end of the filter run.

k
V/ r
300
X. infm«nt turbidity
>

250
>

V
200

y
100
m*/hr
E =0 6 mm
Flow roTe = 4 7 m * / m i / h r 50
E = 0-4mm
•fftu«OT turt>i*ty

6 !2 18 24
Running Tktu { hr* J
effluent rurtoidiry

Fig. 7: Variation of influent turbidity, effluent turbidity and

Running Time t hrs )
head loss with respect to filter run
Fig. 4: Variation of influent turbidity, effluent turbidity and
head loss with respect to filter run

Flow rote i 7 - 1 1 / m * / hr
E = 0 4 mm

ffkjent turbidity
Running Tim* ( firs )
I Time ( hrs )

Variation of influent turbidity, effluent turtridity and Fig. 8: Variation of influent turbidity, effluent turbidity and
head loss with respect to filter run head loss with respect to filter run

PERTANIKA VOL. 12 NO. 1, 1989 115

 AHMAD JUSOH, MEGATJOHARI MEGAT MOHD. NOOR AND ABDUL HALIM GHAZAIJ

\ / " " ^

rote : 7 i n / mV hr
E =O 9«w>

"tr"^ »ff1u«nt turbidity *

o e 12 18 24
Runrara Tim* { mI

Fig. 9: Variation of influent turbidity, effluent turbidity and Fig. 11: Variation of influent turbidity, effluent turbidity and
head loss with respect to filter run head loss with respect to filter run

Flo* rot* : 4 7 mV m*/hr

E = 0 9 mm

Wjp. 70: Variation of influent turbidity, effluent turbidity and

head loss xoith respect to filter run Fig. 12: Variations in total running times at two different flow
rates for different effective sizes.

Figure 12 shows the relationship between
the total filter run and the effective size at the
two different flow rates as mentioned earlier.
The total running time at the higher flow rate
(7.1 mVnr/hr) is shorter than that at the lower
flow rate (4.7 mVm 2 /hr) for the same effective
size. This is as anticipated because at the higher
flow rate, the filter is loaded with a higher sus-
pended solids content in a shorter period of
time. The study shows that an average reduc-
tion of 2 hours in the total filter run occurs as
a result of a 50% increase in the flow rate (from
4,7 mVmVhr to 7.1 mVmVhr) for all cases.
CONCLUSION
The results obtained indicate that the optimum
depth of filter media reduces slightly with the
reduction in effective size. Variation in inlet
turbidity of less than 5 NTU is not a contribut-
ing factor to the final turbidity. A longer filter
run is achieved with higher effective sizes. An

116 PERTANIKA VOL. 12 NO. 1, 1989

 EFFECTS OF EFFECTIVE SIZE IN RAPID SAND FILTRATION

increase in flow rate of 50% resulted in a mini- FAIR, G . M . J . C . GEYERand D.A. OKUN. 1968. Water
mal reduction of the total filter run. and Wastewater Engineering}/o\ume 2, p. 27-4. New
York: John Wiley 8c Sons.
AHMAD JUSOH HAMMER, M.J. 1986. Water and Wastewater Techno-
MEGAT JOHARI MEGAT MOHD. NOOR logy. p.247. New York: John Willey 8c Sons.
ABDUL HALIM GHAZALI SCHULZ, C.R. a n d D.A. OKUN. 1984. Surface Water
Treatments in Developing Countries, p. 146. London:
Department of Civil and John Wiley & Sons.
Environmental Engineering, WHO. 1984, Guidelines for Drinking Water
Faculty of Engineering, Quality, Vol. 1 : Recommendations. World Health
Universiti Pertanian Malaysia, Organization, Geneva.
43400 UPM Serdang, Selangor Darul Ehsan,
Malaysia.

REFERENCES
CLARK, J.W., W. VlESSMAN, Jr. and M.J.
HAMMER. 1977. Water Supply and Pollution Con-
trol. Third Edition, p. 384. New York: IEP. (Received 25 October, 1988)

PERTANIKAVOL. 12 NO. 1, 1989 117

 Section III
Social Sciences
Pertanika 12(1), 119-123 (1989)

Estimating the Income and Substitution Effects

on the Demand for Poultry Meat
ZAINALABIDIN MOHAMED and ROSLAN A. GHAFFAR
Department of Economics,
Faculty of Economics and Management
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

Key words: Poultry meat; demand; Slutsky equation; elasticity.

ABSTRAK
Kertas kerja ini cuba menganilisis dan menentukan faktor-faktor yang mempengaruhi permintaan daging
ayam di Malaysia. Secant Spesifik, ianya cuha melihat perubahan terhadap permintaan daging ayam yang
disebabkan oleh perubahan harga, pendapatan dan juga terhadap perubahan harga barang-barang yang
berkaitan dalarn pasaran daging. Dua model permintaan daging ayam telah diuji, ianya adalah persamaan
tunggal dan persamaan pembolehubah 'instrumental. Di sampingitu juga persamaan Slutsky juga digunakan
untuk menganggar kesan pendapatan dan gantian terhadap daging ayam dibandingkan dengan daging-
dagingyang lain. Analisis menunjukkan bahawa daging ayam adalah barangbiasa dan daging lembu dan
babi adalah barang pengganti bagi daging ayam. Walau bagaimanapun keputusan yang bercampuran di-
dapati di antara hubungan daging ayam dengan ikan. Berdasarkan kepada kesan penggantian dan keanjalan
silang harga, didapati permintaan ikan adalah tidak berkait dengan barang-barang lain.

ABSTRACT
This study attempts to analyse the factors underlying the demand and consumption trends of poultry meat
in Malaysia. Specifically, it explores the poultry meat consumption due to changes in prices and income and
the interaction in demand between poultry meat and the components of the rest of the meat market. To obtain
the parameters of the demand model for poultry meat, two estimation techniques were used in estimating the
demand model, namely single and instrumental variables approach, while Slutsky equations were utilized to
estimate the income and substitution effects ofpoulhy meat with the other components of the meat market. The
results indicate that poultry meat is a normal good while both beef and pork are substitutes to poultry meat.
However, mixed results were obtained for poultry meat and fish. Based on the magnitude of the substitution
effect and cross price elasticity, it is concluded that the demand for fish is independent of the other components
of the meat market.

INTRODUCTION and the components of the rest of the meat

This study attempts to estimate and evaluate
the consumption pattern of poultry meat. Spe-
cifically this paper explores the effect on poultry
meat consumption due to changes in prices
and income of the consuming population. Also
included in the analysis is the nature of the
interaction in the demand between poultry meat
market.
This study is motivated by several reasons.
First, despite the impressive growth of the
Malaysian poultry sector1, a review of the litera-
ture reveals that there is inadequate informa-
tion with regard to the demand structure of
poultry meat and its implications on the rest of

A recent account of the developments in the poultry sector are given in Mohamed et al. (1988).
 ZAINAIABIDIN MOHAMED AND ROSIAN A. GHAFFAR
the red meat subsectors. A few empirical stud-
ies that have been conducted include those by
Ng Yoke Yen (1976) and by Lee Peng Seng
(1980). Considering that these studies are quite
dated, more recent information on poultry meat
demand is warranted.
Poultry meat is one of the several meats
available in the Malaysian market. Questions
pertaining to the substitutability or comple-
mentarity of poultry7 meat with respect to the
other meats (eg. mutton, pork and beef) also
need to be answered. It is reasonable to consi-
der the demad for poultry meat to be less res-
ponsive to price changes as it is acceptable to
all communities in Malaysia. This question, how-
ever, has not been rigourously investigated. An
attempt along this line can be gleaned from the
study done t>y Mohamed et al (1988). The
present study, however, differes in the sense
that it also makes use of the Slutsky equation
to measure the net substitutability/complemen-
tarity of poultry meat with otermeats. In order
to analyse the substitutability or complemen-
tarity effect of poultry meat on other meats, two
models will be presented. The first model is a
linear demand equation for poultry meat.
Parameter estimates from this model are then
used in model II which is basically the Slutsky
equation to obtain the mean substitution effect
between poultry meat and the rest of the meat
products. Thus from such an analysis, the net
substitution or complementarity between
poultry meat and other commodities men-
tioned can be measured.
The rest of this paper is presented in three
sections. The method and data used in this
study will be presented in the following section.
In section three the empirical results and some
discussion on the results will be presented. The
last section provides our summary and conclu-
sion.
THE MODEL SPECIFICATION
Model 1
The demand function for poultry meat can be
written as follows
Q, = f(Pa , Pb , P , Yf , U) (1)
where
120 I'l•RlWIKAVOl., 12 NO. 1. 19SVI
Q = aggregate per capita consumption of
poultry meat (kg)
P = retail price of poultry meat ($/kg)
P* = retail price of beef ($/kg)
P} = retail price of pork ($/kg)
P( = aggregate retail price of fish ($/kg)
Yf = aggregate per capita income ($/year)
U = random disturbance terms
The signs of the variables appearing on
the right hand side are expected to conform to
demand theory. The sign upon own price is ex-
pected to be negative, which implies an inverse
relationship between price and the commodity
consumed. The substitute or competing and
complementary items are expected to have a
direct dnd inverse relationship, respectively, with
the commodity consumed. In this regard the
expected sign is positive and' negative respec-
tively. Lastly, a positive relationship is also ex-
pected between the commodity consumed and
the aggregate per capita income.
A linear additive form of equation was
utilized to analyse the demand function. In this
case, two versions of the demand equation for
poultry meat were formulated and estimated
using two different techniques, namely Ordi-
nary Least Squares (OLS) and the Instrumental
Variables (IV) approach.
Model II
The second model is the comparative statics of
the utility maximization model in the form of
Slutsky equations which can be presented as
where
X. = quantity of the ith commodity demanded
P = price of jth good
Y = per capita income
The alphabets y, u, and p, which appear
as subscript and superscript in equation (2)
indicate income, utility and price, respectively,
and are held constant.
The Slutsky equation can be broken up
conceptually into two parts. The first term on
the right hand side of equation (2) is the pure
substitution effect, or a response to a price
change holding the consumer on the original
 ESTIMATING THE INCOME AND SUBSTITUTION EFFECTS ON THE DEMAND FOR POULTRY MEAT
indifference surface. The second term on the
right is the pure income effect where income
is changed, holding price constant, to reach a
tangency on the new indifference curve. Invok-
ing the envelop theorem (Silberberg, 1978) for
the cost minimization problem, the subtitution
effect can also be written as
In short, equation (3) can be further rewritten
as
6X.11
K.11 = - (4)
SP.
where X., P and Y, and 8*, y and p are defined
as before.
According to Philips (1974) the total effect
of price changes 8X.V/8P] is negative if SX.V/
8Y > 0, or if 8X.V8Y < 0 and 18X; (8X.78Y) f <
I K.. I, which applies for normal good. The total
effect of a change in price, 8X.V8P. can be
positive if 8X.7SY> 0 and 8X.V(8X78Y) < I K.I,
which is the case for Giffen paradox. Hence the
good is superior if 8X.V/8Y > 0, and inferior if
8X.V8Y < 0.
On the other hand as indicated by Hick's
(1946), based on total substitution effect, K..,
goods i and j are classified as substitutes,
complements or independent depending on
whether K is positive, negative or zero, respec-
tively.
Thus given the means of the observed
value of X.(X. = 1, 2 and 3), that is, the three
main variables in equation (1) and the regres-
sion coefficients estimated from the equation,
the substitution and income effect between
poultry meat and other meat can be estimated
through Model II.
Source of Data
Yearly time series data have been used for the
period 1960-1984 and were obtained from
various sources. Consumption data were ob-
tained from the Division of Veterinary Services
(DVS), while data on price were obtained from
various FAMA (Federal Agricultural Marketing
2
Another demand model was estimated for both OLS and IV approaches by dropping the price of Fish from the model.
All variables do conform to demand theory. However the coefficient, R2 and level of significance do not differ from
Model I.
PERTANIKA VOL. 12 NO. 1, 1989
Authority) bulletins, Popultion, income and con-
sumer price index (CPI) were obtained from
various Malaysia Plans and Economic Reports.
All retail prices and income data were deflated
by CPI (1967 = 100). The per capita consump-
tion figures were derived by dividing the total
consumption with the total consuming popula-
tion.
RESULTS AND DISCUSSION
The estimation procedure used for estimating
the demand equation was ordinary least square
(OLS) with Cochrane-Orcuttlnterative Technique
(COIT) to adjust for serial correlation for the
single equation. Since we are also estimating the
aggregate demand equation for poultry meat
in a system of equations, the instrumental vari-
ables technique (IV) were utilized to alleviate
the simultaneity bias. The instrumental variables
choosen in this case are Pb, P^ and P(.
The results for both OLS and IV estima-
tor of model I are presented in Table 1. The
results for the linear equation show that, except
for Pf which has a negative sign, the signs on
Pa, Pb, Pp and Yare met as expected. Although
the expected signs on all varibles, except fish,
do conform to demand theory, statistically, these
variables are not significant at the 10% proba-
bility level. The adjusted coefficient of determi-
nation (R2) is 0.96 which indicates that approxi-
mately 96% of the variation in poultry meat
consumption can be explained by the variation
in the exogenous variables. For the IV estima-
tor, the results in Table 1 show an improve-
ment in the significance of the variables in the
model especially for pork price and per capita
income. However, the adjusted coefficient of
determination was also 0.96'
Thus given the estimated coefficients from
Model I, Model II can be estimated without dif-
ficulty. Table 2 presents the substitution, in-
come and total effect for both OLS and IV
approaches.
The results as presented in Table 2 show
that the total effect of a change in own price
is negative which indicates that poultry meat is
a normal good. While according to Hick's de-
121
 ZAINALABIDIN MOHAMED AND ROSLAN A. GHAKFAR

finition, both poultry meat and beef, poultry indicating that poultry meat is a normal good.
meat and pork are substitutes. However, the Beef appears to be the stronger substitute for
results obtained for poultry meats and fish poultry when compared with pork, which is
generated by the OLS and IV approaches are consistent with the finding in Model II. How-
mixed. In the single equation approach, the ever the cross-elasticity between poultry meat
substitution effect indicates that poultry meat and fish is very inelastic, almost approaching
and fish are substitutes while in the simultane- zero.
ous equation it is a complement. Nevertheless,
the total effect indicates that both poultry meat TABLE 2
and fish are complementary to each other. Estimated substitution, income and total effects
of poultry meat with respect to other meat:
Model II
TABLE 1
Estimated regression coefficients and summary OLS IV
statistics for poultry meat demand equation
Substitution effect (SE)
Variable OLSa IYa
K * -3.9081 - 2.453
Intercept 12.029 1.9603 0.9150

I:
0.7944
0.4091 2.5116
p., - 3.9824 -2.5751
0.2078 -0.1596
(1.0944)* (0.7123)*
p,, 0,8976 0.7907
Income effects (IE)
(0.8962) (0.8192) 0.0739 0.0171
0.0174 0.0040
p 0.4019 2.4948 .1.1

I* I 0.0724 0.0168
(0.5674) (1.0701)
pt - 0.0060 -0.2091 c 0.2138 0.0495

(0.2532) (-0.2488) Total Effect (TE)

-3.9819 -2.4701
Y 0.0095 0.0022 T 0.8976 2.4948
(0.0120) (0.0046)** 0.3367 2.4948
Adjusted R- 0.96 0.96 - 0.006 -0.2091
F - Statistic s L02.50 133.93 *The subscript ab, ap and af indicates that IE, SE and TE
of poultry meat and beef, poultry meat and pork, and
Figures in parentheses are standard errors poultry meat and fish respectively.
* Significant at 5% probability level
** Significant at 10% probability level TABLE 3
Estimates of elasticities of demand
The question now arises as to what level for poultry meat
or magnitude of the substitution effect can indi-
cate the degree of independence of the two Poultry Beef Pork Fish Income
commodities. Nicolaou (1977) suggests that if
OLS - 1.244 0.428 0.138 -0.0014 0.168
the value obtained is around 0.2 and below, it
could be considered as an independent good. IV -0.777 9.369 0.835 -0.047 0.377
Alternatively, according to Johnson, etal (1984)
using the eleasticities one could also tell the On the other hand the own price elastic-
substitutability and complementarity of two ity was inelastic in the case of the IV approach.
goods. Table 3 presents the elasticity coefficients However, pork seems to be a stronger substi-
calculated at the mean of the observed values tute than beef, which is opposite to the OLS
for poultry meat with respect to its own price results. Again the cross-elasticity of poultry meat
(PJ, Ph, Pp, P, and Y for both OLS and IV with respect to fish is very inelastic. Thus from
approaches. For the OLS equation, the own the above discussion one can conclude that fish
price elasticity for poultry meat is elastic at -1.2 is an independent good with respect to poultry

122 PERTANIKA VOL. V2 NO. 1, 1989

 ESTIMATING THE INCOME AND SUBSTITUTION EFFECTS ON THE DEMAND FOR POULTRY MEAT
meat c o n s u m p t i o n . It is to be n o t e d that, the
results p e r t a i n i n g to the substitution effect or
i n c o m e effect, total effect, a n d elasticities de-
p e n d u p o n w h e t h e r the m e a n of X (j = 1, 2
and 3) provide realistic value for X, w h e t h e r
or n o t K is significant, a n d lastly, w h e t h e r or
not the m o d e l is correctly specified.
CONCLUSION
This study is an attempt to estimate the de-
m a n d for poultry m e a t in Malaysia. T h e vari-
ables were selected based on e c o n o m i c theory
and a priori expectation. T h e results indicate
that poultry meat is a normal g o o d while both
beef a n d pork are substitutes to poultry meat.
Mixed results were obtained between poultry
meat a n d fish. It is c o n c l u d e d that fish is an in-
d e p e n d e n t g o o d based on the m a g n i t u d e of
the substitution effect a n d cross price elasticity.
H e n c e the results from Model II a n d the com-
p u t e d elasticity values are generally consistent.
T h e c h a n g e in i n c o m e was also found to
g e n e r a t e little response in poultry meat con-
sumption as indicated by the i n c o m e elasticity.
T h e results seem to be in conflict with the value
of 1.9 a n d 1.02 o b t a i n e d by Ng Yoke Yen (1976)
a n d Lee P e n g Seng (1980) respectively. While
differing sampling periods a n d estimation pro-
cedures may c o n t r i b u t e to the difference, o n e
can also he confortable with the idea that poultry
meat may no longer be a luxury item as it was
twenty years ago. T h u s t h e r e is a great potential
for the expansion of poultry meat c o n s u m p -
tion as it is a p p e a l i n g to every class of t h e
population.
PERTANIKA VOL. 12 NO. 1, 1989
REFERENCES
DIVISION of VETERINARY SERVICES. Various reports.
KAMA. Various issues of FAMA Bulletin.
HICKS, J.R. 1946. Value and Capital, 2nd ed.
Oxford: Oxford University Press.
JOHNSON, R.S, ZUHAIR HASSAN and R.D.
GREEN. 1984. Demand System Estimation Method
and Applications. Ames: The Iowa States University
Press.
LEE PENG SENG. 1980. Demand Analysis of live-
stock Meat in Peninsular Malaysia. Unpublished
Project paper, Universiti Pertanian Malaysia.
MAIAYSIA. Ministry of Finance. Economic Reports,
various issues.
MALAYSIA PLANS. Prime Minister Department. Vari-
ous issues.
No YOKE YEN. 1976. Demand Analysis of the Live-
stock Meats and Projection of the Demand for
Livestock Meats and Feedstuffs in West Malaysia.
Unpublished Project Paper, University of Malaya.
NlCOLAOU, G.B. 1977. The Place of Petroleum
in the U.K. Fuel Market. Applied Economics, 9: 167-
172.
PHILIPS, I,. 1974. Applied Consumption Analysis.
Amsterdam: North Holland.
SlLBERG. 1978. The Stru ctu re of Economics: A Mathe-
matical Analysis. McGraw-Hill Book Company.
ZAINALABIDIN MOHAMED, MAD NASIR SHAMSUDIN and
ROSLAN A. GHAFFAR. 1988, Staff Paper No. 2,
A Study of Demand for Poultry Meat in Malaysia.
Department of Agricultural Economics, Univer-
siti Pertanian Malaysia.
(Received 27 July, 1988)
123
Pertanika 12(1), 125-135 (1989)

Community Leader Participation in the

Extension Programme Decision-Making Process
AMINAH AHMAD
Centre for Extension and Continuing Education
Universiti Pertanian Malaysia
43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

LAVERNE B. FOREST
Department of Continuing and Vocational Education
University of Wisconsin
Madison, WI 53705, USA

Key words: Adult leaner particpation; program development; decision-making process.

ABSTRAK
Penyelidikan ini mengkaji taraf penyertaan ketua masyarakat dalam proses membuat keputusan bagi program
pengembangan, berdasarkan tanggapan pegawai pengembangan. Objektif kajian adalah untuk menentukan
(1) sama ada taraf penyertaan berbeza mengikut jenis keputusan yangperlu dibuat bagi dua program yang
berlainan, (2) sama ada taraf penyertaan berbeza mengikut langkah-langkah proses membuat keputusan, dan
(3) bagaimana kaitan taraf penyertaan dengan sikap pegawai pengembangan terhadap penyertaan ketua
masyarakat bagi tiap-tiap jenis program pengembangan. Penyelidikan ini mendapati (1) berlawanan dengan
jangkaan, taraf penyertaan ketua masyarakt tidak berbeza bagi kedua-dua jenis program 'lebih teknikaV dan
'kurang teknikal', (2) taraf penyertaan berbeza mengikut langkah-langkah proses membuat keputusan. Ketua
masyarakt paling kurang menyertai di langkah-langkah pemilihan alternatif dan perancangan tindakan.
(3) ada kaitan positif di antara tanggapan pegawai pengembangan dengan taraf penyertaan ketua masyar-
akat, hanya bagi program 'kurang teknikal.'

ABSTRACT
This research studied participation levels of community leaders in the extension programme decision-making
process, as perceived by extension staff The objectives were to determine (1) if the perceived particpation levels
differed by tasks associated with two different types of programme decisions, (2) if the perceived participation
levels varied by tasks associated with different steps of the decision-making process, and (3) how the perceived
participation levels relate to extension staffs attitudes toward community leader participation, for each type
of programme decision. This research found the following: (1) Contrary to expectation, the perceived partici-
pation levels of community leaders did not differ for both "more technical' and 'less technical7 programme
decisions: (2) The perceived participation levels varied by the different decision-making steps. Community
leaders were perceived to participate least at the choice of alternatives and the plan of action steps of decision
making. (3) Only for 'less technical' programme decisions were attitudes of extension staff positively related
to perceived community leader participation levels.

INTRODUCTION and Van de Ven (1980), adult learner partici-

Literature on adult education programme de- pation helps (1) ensure acceptance and sup-
velopment emphasises the importance of adult port of the resulting programs; (2) make more
learner participation in programme develop- accurate decisions about adult learners' needs,
ment. According to Boyle (1981), Forest (1976), interests, and problems; and (3) inform and
 AMINAH AHMAD AND LAVERNE B. FOREST
prepare adult learners for active leadership roles
in the future.
Realising the importance of participation
in this process, extension organisations and staff
have generally directed their efforts towards
increasing adult learner participation in pro-
gramme decision making. Their efforts could
be more successful if they had an increased un-
derstanding of factors influencing the levelof
adult learner participation. Among the factors
often neglected by extension organisations,
when involving clientele and leaders in the
planning process, are the nature of decision
tasks, and attitudes of extension staff guiding
the participation. The decision tasks are asso-
ciated with different types of programme deci-
sion and with different decision-making steps.
This research examined (1) the difference
in FELDA community leader participation levels
in two different types of programme decisions,
as perceived by extension staff, (2) differences
in participation levels in six identified decision-
making steps, as perceived by FELDA exten-
sion staff, (3) relationships between the per-
ceived community leader participation levels
and extension staff attitudes toward community
leader participation in decision making, for
two different types of programme decisions.
Theory and Hypotheses
In adult education programme decision ma-
king, both adult educators and adult learners
face different types of decision tasks. The deci-
sion tasks can be associated with the type of
programme decisions (Maier 1963), and the
tasks can be associated with the process of deci-
sion making which consists of several steps
(Last 1972).
The literature has frequently described
two categories of programmes dealing with
different decision types. The first category re-
quires relatively more technical information
for decision making, such as a programme on
agriculture production. Decisions associated
with such a programme have been described
as routine decisions by Delbecq (1967), high-
quality decisions by Maier (1963), closed deci-
sions by Jimmerson (1981), programmed deci-
sions by Simon (1960), and computational deci-
sions by Thompson (1957). The second pro-
gramme category requires relatively less techni-
126 PERTANIKA VOL. 12 NO. 1, 1989
cal information for decision making, such as a
programme on community development. Deci-
sions associated with such a programme have
been described as creative decisions by Delbecq,
high-acceptance decisions by Maier, open de-
cisions by Jimmerson, heuristic decisions by
Simon, and judgmental decisions by Thompson.
The process of decision making involves
different decision tasks at the different steps.
The decision-making steps can be classified in-
to two categories. Last (1972) suggested a cate-
gory of steps requiring relatively more fact-
oriented information, such as the identifica-
tion of alternatives and evaluation of alternativs
steps. The other suggested category of steps re-
quires .relatively more value-oriented informa-
tion, such as the identification of problem and
choice of alternative steps.
According to the literature on group problem
solving, the type of decision task could effect
the group process, such as participation by
group members in programme decision ma-
king. The group members will adjust their
patterns of communication and roles accord-
ing to the type of task (Delbecq 1967). Com-
munity leaders involved in group decision ma-
king, for example, tend to vary in their partici-
pation as they face different decision tasks.
Besides being affected by the type of de-
cision task, community leader participation in
programme decision making is related to ex-
tension staff attitudes toward adult learner
participation. According to Katz and Kahn
(1978), persons interacting in a role set, in an
organisation, are involved in a cyclic process of
role sending and role receiving. The role sender
attitudes, expectations and perceptions are re-
lated to the role receiver behaviour. In an ex-
tension organisation, extension staff, guiding
community leader participation in programme
decision making, and community leaders parti-
cipating in programme decision making, are
involved in role sending and role receiving.
Based on this role theory, community leaders
behave or participate in reponse to extension
staff attitudes, expectations and perceptions,
and vice versa.
Literature on group dynamics suggests
that group member participation in decision
making could be influenced by the leader or
 COMMUNITY LEADER PARTICIPATION IN THE EXTENSION PROGRAMME DECISION-MAKING PROCESS
the staff guiding the participation. In a study of
managerial leadership styles in problem- solv-
ing conferences, autocratic leadership style did
not encourage group member participation
(Delbecq 1965). Since a person's behaviour
influences his or her attitude toward the be-
haviour (Fishbein and Ajzen 1975), a leader
exhibiting autocratic leadership behaviour is
likely to possess a less favourable attitude toward
group member participation. Hence, extension
staff with less favourable attitudes toward par-
ticipation are not likely to encourage commu-
nity leader participation in programme deci-
sion making.
The literature suggests (1) adult learners
face different types of decision tasks during pro-
gram decision making, (2) adult learner par-
ticipation could differ as they face different
decision tasks, and (S) a relationship possibly
exists between attitudes of adult educators and
the participation of adult learners in programme
decision making. The research reported here is
relative to four hypotheses derived from this
literature.
H1: Perceived community leader participation
levels differ for two different programme
decision types.
H2: Perceived community leader participation
level differs in each of the decision-making
steps, during programme decision ma-
king.
HS: The more positive the extension staff atti-
tudes toward community leader participa-
tion in making 'more technical' program-
me decisions, the higher the perceived
community leader participation levels.
H4: The more positive the extension staff atti-
tudes toward community leader participa-
tion in making 'less technical' programme
decisions, the higher the perceived com-
munity leader participation levels.
MATERIALS AND METHODS
Study Setting, Sampling, and Procedure
Eighty-six FELDA scheme managers (extension
staff) and 94 FELDA settler leaders (commu-
nity leaders) constituted the study samples.
These persons were selected as the study sample
because they met the requirements needed for
the study hypotheses. Monthly Settler Develop-
PERTANIKA VOL.. 12 NO. L, 1989
ment Committee (SDC) meetings chaired by
scheme managers and attended by settler lead-
ers represented the adult educator-adult learner
programme decision-making setting.During
these monthly meetings, the managers expect
the settler leaders to participate in on-going
programme decisions and learn decision-
making skills, so that eventually a settler leader
can take over the chairman role held by each
manager at these meetings.
Furthermore, during these monthly
meetings, managers and settler leaders have
been involved in two broad decision categories
- agriculture production-related and commu-
nity development-related decisions. Decision
making involving these categories represented
the two types of programme decisions studied.
These decisions were studied because they pro-
vided variation in the nature of information
required to make decisions. Agriculture pro-
duction-related decisions need relatively more
technical information to arrive at solutions
compared to community development-related
decisions.
Using two sets of instruments, data were
collected from managers and settler leaders.
Self-administered questionnaires were used to
collect data from managers. The managers pro-
vided data on their attitudes toward settler
leader participation in the programme deci-
sion-making process, and perceived settler
leader participation levels. Using interview
schedules, data on actual participation levels
were obtained from settler leaders. The data
from settler leaders were used to validate the
measurement of perceived participation levels.
Attitude Measurement
This btudy measured manager attitudes toward
settler leader participation in the six rational
steps of decision making identified by Cart-
wright and Zander (1960), Patton and Giffin
(1973), and Phillips (1973). These steps also
constituted Dewey's (1933) reflective thinking
process steps, namely (1) identification of prob-
lem, (2) analysis of problem, (3) identification
of alternatives, (4) evaluation of alternatives,
(5) choice of alternative, and (6) development
of plan of action.
Using the Likert five-point scale method
manager attitudes were measured. Managers
127
 AMINAH AHMAD AND LAVERNE B. FOREST
indicated their levels of agreement toward state-
ments of possible beliefs and intentions about
settler leader participation in each of the six
decision-making steps. Four items for each de-
cision-making step gave a total of 24 items for
an attitude scale.
Two 24-item scales were used to measure
the managers' attitudes, one scale for each of
the two programme decision types i.e. agricul-
ture production-related and community devel-
opment-related decisions. For each scale, the
24 items were selected from 42 pretested items.
For each decision-making step one half of the
items indicated a favourable or positive atti-
tude, the other one half of the items indicated
an unfavourable or negative attitude toward
settler leader participation in the step.
Level of Participation
This study measured the perceived level of
participation in the decision-making process in
terms of a continuum, from a high to a low
level of sharing of information, opinions and
experiences that can influence the decision-
making process. Two scales were developed for
each of the two programme decision types.
The perceived participation levels were
measured by the frequency with which settler
leaders actually performed observable verbal
communication (such as telling about experi-
ences, explaining situations, providing opinions,
and suggesting ideas), in the decision-making
process, during the most recent meeting. State-
ments describing the verbal communication
for each of the six identified decision-making
steps were formulated. Using the Likert five-
point scale method, four items developed for
each step gave a total of 24 items for each scale.
Besides the perceived participation levels,
settler leader self-reported participation levels
were also measured using five-point Likert scales.
Six items, representing the six decision-
making steps, were selected from the 24-item
scale used for measuring the perceived partici-
pation level. Two six-item scales were used to
measure the self-reported participation level,
one scale for each of the two programme deci-
sion types.
Reliability Test of Scales
To calculate the scale reliability coefficients,
the Guttman Model (Guttman 1945) was used.
128 PERTANIKA VOL. 12 NO. 1, 1989
The model was chosen because the attitudes
and participation levels were measured using
composite scales of a simple sum of scores across
items. Using this model, six reliability coeffi-
cient estimates were provided for each scale.
According to Hull and Nie (1981) all estimates
would underestimate the true reliability, and
strategically one should pick the largest of the
six coefficients computed since one could be
sure that the true reliability would be higher.
The reliability coefficients of the scales for the
study are as shown in Table 1.
TABLE 1
Scale reliability coefficients
Scale Reliability
• coefficient
Perceived level of participation
in agriculture production-related
decisions .922
(24 items)
Perceived level of participation
in community development-related
decisions .936
(24 items)
Attitude toward participation
in agriculture production-related
decisions .724
(24 items)
Attitude toward participation
in community development-related
decisions .798
(24 items)
Self-reported participation level
in agriculture production-related
decisions .764
(6 items)
Self-reported participation level
in community development-related
decisions .791
(6 items)
Data Analysis
Using the mean perceived settler leader parti-
cipation level score for each of the two partici-
pation scales, the paired t-test was used to test
the hypothesis of difference between the per-
ceived settler leader participation levels for
the two programme decision types. Analysis of
 COMMUNITY LEADER PARTICIPATION IN THE EXTENSION PROGRAMME DECISION-MAKING PROCESS

variance tested the difference in the perceived process for both programme decisions. Also,
participation levels for the six decision-making settler leaders might be expecting managers
steps. Pearson correlation was used to test the to assume more important roles in the deci-
hypotheses on the relationships between ma- sion-making process for both programme deci-
nagers' attitudes and the perceived settler leader sions. Possibly settler leaders believe that mana-
participation levels, for each of the programme gers have a legitimate right to provide informa-
decision types. tion during the SDC meetings, by virtue of the
position the managers occupy. Also the mana-
RESULTS AND DISCUSSION gers were viewed as the acknowledged experts
HI: Perceived Community leader Participation levels and capable of taking all the necessary factors
Differ for Two Different Programme Decision Types. into consideration during the decision-making
Table 2 shows the means and standard devia- process in the meeting.
tions of perceived participation level and atti-
tude scores. Table 3 shows the t values for the TABLE 2
Means and standard deviations of attitude and
difference in the participation levels between
participation level scores
the two programme decisions. The data did not
support hypothesis 1. No significant difference Participation/Attitude Mean S.I).
was found between the mean perceived parti-
cipation levels for the two programme deci- Attitude toward an interactive
sions - agriculture production-related and role for agriculture production-
related decisions 76.03 1 6.63
community development-related decisions. The
(N = 85)
expectation that the mean perceived participa-
tion level for agriculture production-related Perceived participation level
decisions would differ from that for community for agriculture production-
development-related decisions was not support- related decisions 73.0411 10.89
ed. This study indicated that the nature of the (N = 85)
decision tasks associated with the different pro- Attitude toward an interactive
grammes did not influence the perceived adult role for community-related
learner participation levels in programme de- decisions 79.07* 7.36
cision making. (N = 85)
However, considering the 12.SI average Perceived participation level
number of years the settler leaders had been for community-related
programme participants, the settler leaders did decisions 73.8211 11.49
not show very high perceived participation levels (N - 85)
in programme decision making for either type Self-reported participation
of programme decision. level for agriculture
The need for independence factor, pro- production-related decisions 22.32 b 3.18
posed by Vroom (1960) might provide some (N= 94)
explanation for the results that did not show Self-reported participation
higher participation level in community deve- level for community-related
lopment-related decisions requiring less tech- decisions 24.52 b 2.83
nical information, compared to agriculture pro- (N = 94)
duction-related decisions requiring more tech- 1
Based on 24-item scale
nical information. Possibly the settler leaders h
Based on 6-item scale
had weak independence need and were little
affected by the opportunity to participate. Another possible reason is that settler
Related to the need for independence are leaders generally did not have the adequate
the settler leader expectations. The settler skills and information for greater participation
leaders might be expecting managers to play in decision making. This might especially be so
an equally important role in the decision-making in situations where settler leader educational

PERTANIKA VOL. 12 NO. 1, 1989 129

 AMINAH AHMAD AND IAVERNE B. FOREST

levels were low and where settler leaders lacked in Tables 4 and 5 show significant differences
professional training. in the perceived participation levels of settler
leaders in the different decision-making steps
TABLE 3 for agriculture production-related and commu-
T-value for the difference in participation nity development-related decisions, respectively
levels in programme decision making (p < .05).
Participation level t p
TABLE 4
Perceived level ANOVA for perceived participation levels in
decision-making steps for agriculture
Between agriculture production production-related decisions
and community development-
related decisions -0.38 0.38 Sources SS df MS
Self- reported lex>rl
Between people 1612.86 83 19.43
Between agriculture production Within people 1330.00 420 3.17
and community development-
related decisions - 6.57 0.00 Between
measures 104.29 5 20.86 7.06 0.00
Residual 1225.71 415 2.95
Although managers perceived that settler
leader participation for the two types of pro- Total 2942.86 503 5.85
gramme decisions did not differ, the self-
reported participation levels differed. As shown
in Table 2, the mean participation level in TABLE 5
community development-related decisions was ANOVA for perceived adult learner participation
significantly higher than that for agriculture levels in decision-making steps for
production-related decisions. This difference community development-related decisions
was found to be significant (p < .01, Table 3).
Sources SS df MS
The results indicated that the nature of the
decision tasks associated with the different pro- Between people 1809.19 82 22.06
grammes did influence the perceived adult
learner participation level in programme deci- Within people 1101.67 415 2.65
sion making. It is possible that settler leaders Between
have relatively inadequate skills and informa- measures 49.39 5 9.88( 3.85 .00
tion for participating in agriculture production- Residual 1052.28 410 2.57
related decisions than in community develop- 2910.86 497 5.86
Total
ment related decisions. However, conclusions
were made based on analysis of perceived
participation level data because the scales used A Scheffe post hoc comparison between
to measure the perceived participation levels means test on the participation levels in the
had higher reliability coefficients than those six decision-making steps for each programme
used to measure the self-reported participation decision found that the group mean for the
levels. choice of alternatives and the plan of action
steps differ significantly (p < .05) from the mean
H2: Perceived Community Leader Participation I^vel for another group of steps — identification of
Differs in Each of the Decision-making Step during alternatives and evaluation of alternatives.
Programme Decision Making. The data in this study setting did not
This hypothesis was tested using perceived support the suggestion by Last (1972) that, in
participation levels in the six decision-making what he calls, fact-oriented decision-making
steps, for each of the two types of programme steps, such as the identification of alternatives
decisions — agriculture production-related and and evaluation of alternatives, the participation
community development-related decisions. Data of adult learners would be less than in, what he

130 1>ERTAN1KA VOL. 12 NO. 1, 1989

 COMMUNITY LEADER PARTICIPATION IN THE EXTENSION PROGRAMME DECISION-MAKING PROCESS

calls value-oriented steps, such as the identifi- the managers about their problems with the
cation of problem and choice of alternative hope that the managers will solve the prob-
steps. In this study, settler leaders participated lems, than to decide on solutions. Associated
least in the choice of alternaties and the plan with the solutions are the resources needed to
of action steps, as shown in Table 6. plan for action. The power to decide on re-
source acquisition and allocation could possi-
TABLE 6 bly be in the hands of managers. This could
Mean scores for manager's perceived reinforce the settler leaders to let the managers
participation levels in decision-making steps play more important roles in deciding on the
for agriculture production and community
best alternative to solve their problems, and in
development-related decisions
planning for action.
Decision H3: The More Positwe the Extension Staff Attitudes
toward Community Leader Partidpation in Alaking
Decision-making Agriculture Community
steps production development 'More Technical' Programme Decisions, the Higher
the Perceived Community Leader Participation lj>veLs.
Mean S.D. Mean S.D. Results of correlation analysis are presented in
Table 4. For agriculture production-related
Identification decisions, the result shows no significant corre-
of problem 12.23 2.54 12.77 2.33
lation between manager attitudes and the per-
Analysis of ceived participation levels of settler leaders in
problem 12.42 2.73 12.33 2.39 decision making. This indicates that manager
attitudes did not seem to relate to the partici-
Identification
of alternatives 13.10 2.07 12.53 2.62 pation behaviour of settler leaders. The result
indicated that the attitudes of managers were
Evaluation of not good predictors of the settler leader partici-
alternatives 11.98 2.61 12.18 2.77 pation levels, in situations where programme
Choice of decisions required more technical and objec-
alternatives 11.77 2.17 11.83 2.29 tive information to arrive at a solution.
Plan of action 11.80 2.33 12.00 2.28 According to the role episode model
developed by Katz and Kahn (1978), the role
senders expectations, including personal atti-
The difference in the nature of informa- tude and perception, could influence the role
tion required at the different steps involving behaviour of the role receiver. In this study,
different decision tasks did not provide an the role senders' expectations were indicated
explanation. The results could possibly be due by the managers' attitudes toward an interac-
to the need for the independence factor as tive role or toward participation of settler lead-
suggested by Vroom (1960), It could be argued ers in the decision-making process for agricul-
that the settler leaders' need for independence ture production-related decisions. The role be-
in making decisions might still be low. Although haviour was indicated by settler leaders partici-
settler leaders participated in all decision- pation as perceived by the managers, in the
making steps, the results suggested they still decision-making process for agriculture produc-
depended on managers in choosing the best tion-related decisions.
alternatives to solve their problems. Another The results indicated that the role send-
possible factor is the settler leaders' acknow- ers* expectations did not significantly influence
ledgement of managers' legitimate right to the role behaviour of the role receivers. This
provide information and decide on the best could be due to some internal sources of
alternative to solve their problems, by virtue of motivation within the settler leaders that led
the position the managers occupy. them to take on a role behaviour that is not in
During the monthly meetings, it is likely accordance with the managers' expectations.
that settler leaders are more eager to inform According to Katz and Kahn (1978) each person

PERTANIKAVOL. 12 NO. 1, 1989 131

 AMINAH AHMAD AND IAVERNE B. FOREST

could be a 'self-sender', that is, a role s e n d e r
to himself or herself. In this study, the role
receiver h a d a c o n c e p t i o n of his or h e r position
as a settler leader, a n d a set of attitudes a n d
beliefs a b o u t what he or she should a n d should
n o t d o , with regards to participation in pro-
g r a m m e decision making. In o t h e r words, the
settler leaders' own cognitions or expectations
could con-tribute to the absence of correlation
between the two variables. If this is so, t h e n t h e
mana-gers' and settler leaders' expectations were
n o t congruent. This, according to Katz and Kahn
(1978), is called intersender role conflict. T h e
role sent by settler leaders themselves was in
conflict with that sent by the managers.
H4: The More Positive the Extension Staff Attitudes
toward Community Leader Participation in Making
'Less Technical*Programme Decisions, the Higher the
Perceived Community Leader Participation Levels.
Data in Table 7 show a significant positive cor-
relation between manager attitudes and per-
ceived perticipation levels when making com-
munity development-related decisions. The
correlation coefficient of 0.24, although small,
was statistically significant (p < .01).
TABLE 7
Relationships between perceived participation
levels and attitudes toward community leader
participation in programme decision making
In situations where settler leaders culti-
vated oil-palm, all relationships had higher
correlation coefficients (Table 8) than those
reported in Table 7. The magnitude of the
difference between the correlation coefficients
was greater for comunity development-related
than for agriculture production-related deci-
sions. The r value for community development-
related decisions was found to be higher
(r = .37, Table 8) compared to that reported in
Table 7 (r = .24). The r value for community
development-related decisions value would
more likely be higher with a greater sample
number in situations where oil-palm was culti-
vated.
The relationship between overall manager
attitudes toward settler leader participation and
overall perceived participation levels was sig-
TABLE 8
Relationships between attitudes and perceived
participation levels in programme decision
making, according to the type of crop cultivated
Relationship N r P
Oil-palm
Attitudes and perceived
participation levels for
agriculture production-
related decisions .15 .15

Relationship r p Attitudes and perceived

participation levels for
Attitudes and perceived community-related
participation levels for decisions 34 .37 .01
agriculture production- Overall attitudes and
related decisions .05 .32 overall perceived
Attitudes and perceived participation levels M .25 .04
participation levels for Rubber
community development
related decisions .24 .01 Attitudes and perceived
participation levels for
Overall attitudes and agriculture production-
overall perceived related decisions 49 -.21 .11
participation levels .14 .10
Attitudes and perceived
participation levels for
The low correlation value could possibly community-related
be due to other moderating variables affecting decisions 48 .04 .41
the relationship. Further data analysis accord-
Overall attitudes and
ing to, the crop cultivated, showed that the type
overall perceived
of crop cultivated had an influence on the
participation levels 48 -.10 .29
relationship.

132 PERTANIKA VOL. 12 NO. 1, 1989

 COMMUNITY LEADER PARTICIPATION IN THE EXTENSION PROGRAMME DECISION-MAKING PROCESS
nificant (p < .05) with an r value of .25, in
situations where settler leaders cultivated oil-
palm. The type of crop had a considerable in-
fluence on the studied relationship. The results
supported Hypothesis 4 for schemes in which
managers and settler leaders were involved with
oil-palm management and cultivation. This
result suggests that the overall attitude of mana-
gers toward settler leader participation influ-
ence the overall perceived perticipation level of
settler leaders in programme decision making,
in situations where settler leaders cultivated oil-
palm.
In situations where settler leaders culti-
vated rubber, the correlation coefficients were
negative for agriculture production-related
decisions and for overall decisions. The coeffi-
cient decreased for community development-
related decisions. However, all the relationships
were non-significant (see Table 8).
The differences in results obtained from
analysing the data according to the type of crop
could be explained by the difference in the
way farming activities are carried out for the
two crops. For oil-palm, group farming is imple-
mented through the 'block system/ In this
system, a group of 20 settlers work on a 200-
acre plot or 'block' of land. The settlers are
collectively responsible for carrying out farm-
ing activities which include fertilizer applica-
tion, weed, disease and pest control, harvesting
of crops, transportation of harvests to the fac-
tory, and keeping farm accounts. For rubber,
these farming activities are all individually
done by the settlers.
Therefore, there is a greater need for
group cooperation in carrying out farming ac-
tivities for oil-palm than for rubber which leads
to a greater need for group participation in
decision making. This is especially so for the
harvesting and transportation of fresh oil-palm
fruit bunches. Unlike oil-palm, the collecting
and transportation of the rubber produce is
done individually.
An analysis of the operations in managing
oil-palm and rubber crops suggests how the
nature of task processes provide the foundation
upon which settler community social structures
are built. According to Perrow (1967), and Hage
and Aiken (1969), technology or the nature of
PERTANIKAVOL. 12 NO. 1, 1989
organisational tasks could influence the nature
of the organisational social structure. That is,
technology likely determines whether there is
a diverse or relatively simple division of labour.
These division of labour patterns could set limits
on the goals that are maximized and those
that are minimized. Viewing the settlers as
members in a sub-social system, the nature of
the task processes in which they are involved
could determine the need for interaction and
the need for expertise. This could in turn deter-
mine the extent of settler leader participation
in the programme decision-making process.
The nature of the task processes could explain
the difference in the results obtained for the
two crops.
The group farming system, implemented
through the block system, was suspected to in-
crease both scheme management and commu-
nity development decision making among oil-
palm scheme settler leaders, at the block level.
It could be reasoned that this experience in
group decision making at the block level
coupled with the manager attitudes and expec-
tations, influenced the settler leader overall
participation level in the decision-making
process at the SDC meeting.
CONCLUSIONS
First, this study concluded that in situations
where community leaders' educational levels
were low and community leaders lacked profes-
sional training and past participation experi-
ences, the perceived participation levels in
making both 'technical1 and 'less technical1
programme decisions were the same and were
not very high.
Second, the perceived community leader
participation levels were different for the
different decision-making steps, for both pro-
gramme decisions. Community leaders partici-
pated least on the choice of alternatives and the
pain of action steps.
Third, the attitudes of -extension staff
toward community leader participation was not
related to perceived participation levels in de-
cision making, for programme decisions requir-
ing more technical information to arrive at
solutions. Only for less technical programme
decisions were more favourable extension staff
133
 AMINAH AHMAD AND IAVERNE B. FOREST
attitudes likely to relate more strongly to com-
munity leader participation levels.
Fourth, the type of programme has a
moderating influence on the relationship bet-
ween extenstion staff attitudes or leadership
orientations, and community leader participa-
tion levels.
Implications for Practice
This study has implications for administrators
of adult education programme development. If
as determined, the type of programme deci-
sions, and adult educator attitudes toward adult
learner participation, are important factors in-
fluencing adult learner participation in pro-
gramme decision making, the type of pro-
gramme decision should be identified when
advocating adult learner participation. Pro-
gramme decisions requiring less technical
information, as in community development-
related decisions, need adult educators with
more democratic leadership orientations and
processes. For programme decisions needing
more adult learner interaction to achieve group
objectives, adult educators with more favour-
able attitudes toward participation are more
likely to enhance adult learner participation in
programme decision making.
It is therefore important for FELDA to
train managers in leadership orientations and
styles suited to different programme decisions.
FELDA managers should be more flexible in
dealing with different programme decisions.
When dealing with community development-
related decisions, more democratic leadership
orientations are needed in order to encourage
settler leader participation in programme
decisions, than with agriculture production-
related decisions.
This study has implications for FELDA
which has as its ultimate goal, the handing over
of decision-making responsibilities to settler
leaders. The fact that settler leaders had been
in the schemes for an average of 12.31 years
leads one to believe that FELDA would expect
them to achieve higher participation levels in
decision making regarding scheme manage-
ment and settler community development.
However, as indicated earlier, the parti-cipa-
tion levels were not very high. This could
especially be so in situations where settler leaders
134 PERTANIKA VOL. 12 NO. 1, 1989
did not have the necessary information and
ability for making the decisions.
There is therefore a need for FELDA to
develop the settler leader management and
decision making skills, and leadership abilities
through a planned training programme with
more specific objectives within an anticipated
period of time. Managers need to expand settler
leaders' responsibilities in management and
decision making in keeping with settler lead-
ers' developing abilities and increasing know-
ledge in management and decision making,
such that in every scheme, one of the settler
leaders will eventually be able to take over the
scheme manager's role as the chairman of the
Settler Pevelopment Committee.
In addition to training and expanding
settler leaders' responsibilities, 'FELDA should
provide settler leaders with appropriate incen-
tives in terms of monetary rewards for the
services rendered by the leaders. This would
encourage settler leaders to participate more
effectively and provide continued leadership in
scheme management.
In the light of these findings, FELDA
managers should be aware of their attitude
influence on settler leader participation for
community development-related decisions. The
findings also suggest the importance of the per-
sistence of this favourable attitude toward set-
tler leader participation among FELDA staff
interacting with settler representatives, at the
different decision-making levels in the organ-
isation.
Implications for Further Research
The relationships between adult educator
attitudes and perceived adult learner participa-
tion levels, obtained from analysis of the oil-
palm study sample (N = 34) data, were stronger
than those found in the rubber sample (N =
48). This finding suggests the need to improve
the research design. Instead of studying adult
educators and adult learners in two settings,
requiring relatively different levels of interac-
tion, as in oil-palm and rubber settings, in the
future the study sample should include only re-
spondents in a decision-making setting requir-
ing relatively more group interaction to plan
and implement adult learner programmes, as
in the oil-palm setting.
 COMMUNITY LEADER PARTICIPATION IN THE EXTENSION PROGRAMME DECISION-MAKING PROCESS
A follow-up study is n e e d e d to measure
adult l e a r n e r participation levels as r e p o r t e d by
adult learners themselves, using a 24-item scale
similar to that used for measuring adult edu-
cators' perceived levels of adult learner parti-
cipation, instead of the 6-item scale. T h e in-
creased n u m b e r of items would be m o r e likely
to improve the reliability of the self-reported
scale a n d would further validate the participa-
tion level m e a s u r e m e n t . T h e data obtained
using this scale could be used to test the hypo-
thesis of relationship between adult educators'
attitudes a n d adult learners' actual participa-
tion levels.
Studies should be c o n d u c t e d on variables re-
lated to adult learners, including adult learn-
ers' n e e d for i n d e p e n d e n c e a n d expectations.
T h e r e is also a n e e d to study variables related
to the cultural context such as the adult learn-
ers' p e r c e p t i o n s on hierarchical relationships,
power a n d authority that could influence the
e x t e n t of adult learner participation.
REFERENCES
BOYLE, P.G. 1981. Planning Better Programs. \Vw
York: McGraw-Hill.
CARTWRIGHT, D. and A. ZANDER. I960. Group Dyna-
mics. Evanston, Illinois: Row, Peterson & Co.
DEWEY, J. 1933. Mow We Think. New York: P.C.
Heath & Co.
DELBEGQ, A.L. 1967, The Management of Deci-
sion Making within the Firm: Three strategies for
Three Types of Decision Making. Academy of
Management Journal 10: 329-339.
DELBECQ, A.L. 1965. Managerial Leadership Styles
in Problem-solving conferences II. Academy of Ma-
nagement Journal 8: 32-45.
FlSHBEIN, M. and I. AjZEN. 1975. Belief Attitude,
Intention and Behaviour: An Introduction to Theory
and Practice. Reading, Massachusetts: Addison-
Wesley.
FOREST, L.B. 1976. Shazvano County leaders: Involve-
ment in Extension Program Development. Madison:
University of Wisconsin - Extension, Publication
No. 6.
PERTANIKA VOL. 12 NO. 1, 1989
GUTTMAN, L. 1945. A Basis for Analyzing Test-
Retest Reliability. Psychometrika. 10: 255-282.
HAGE, J. and M. AlKEN. 1967. Program Change
and Organizational Propertices, A Comparative
Analysis. American Journal of Sociology. 72: 503-510.
HULL, C.H. and N.H. NlE. 1981. SPSS Update 7-9.
New York: McGraw-Hill.
JlMMERSON, R.M. 1981. What Determines Citizen
Involvement? journal of Extension September/
October: 21-26.
KATZ, E. and R.L. KAHN. 1978. The Social Psycho-
logy of Organization (2nd ed). New York: Wiley.
lAST, D. 1972. Principles and Practice of Citizen
Involvement with Special Preference to Regional
Planning in Dane County, Wisconsin. Unpub-
lished Masters Thesis, University of Wisconsin,
Madison.
MAIER, N.R.F. 1963. Problem Solving Discussions and
Conferences. New York: McGraw-Hill.
PATTON, B.R. and K. GlFFlN. 1973. Problem-Solving
Group Interaction. New York: Harper 8c Row.
PERROW, C. 1967. A Framework for the Com-
parative Analysis of Organization. American Socio-
logical Review. 32(2) April: 194-208.
PHILLIPS, G.M. 1973. Communications and the Small
Group. New York: Bob's Merrill.
ROTHMAN, J. 1979. Three Models of Community
Organization Practice, Their Mixing and Phas-
ing. In Strategies of Community Organization F. ed.
M. Cox, J. L. Erlich, J. Rothman a n d j . E. Trop-
man. Itasca, Illinois: F. E. Peacock Publishers.
SIMON, H.A. & A. NEWELL. 1958. Heuristic Prob-
lem Solving: the Next Advance in Operations Re-
search. Operations Research Journal Jan.-Feb.: 1-10.
THOMPSON, J.D. 1967. Organization in Action. New
York: McGraw-Hill,
VAN DK VEN, A.H. 1980. Problem Solving, Plan-
ning and Innovation, Part I: Test of the Program
Planning Model. Human Relations. 33: 771-740.
VROOM, V.H. 1960. Some Personality Determinants of
the Effects of Participation. Engelwood Cliffs, New
Jersey: Prentice-HalL
(Received 4 August, 1988)
135
Agronomy Journal
The American Society of Agronomy's bi-
monthly journal is highly respected for its
technical reports on all aspects of plant
and soil science. Original research find-
ings in crop science; soil science; biome-
try; turfgrasses; agroclimatology and agro-
nomic modeling; crop, soil, pasture, and
range management; and crop, forage, and
pasture production and utilization are
presented.
Letters, invited reviews and notes on ap-
paratus, experimental techniques, and ob-
servations/preliminary results and com-
puter programs are also presented.
$65.00 per year or $13.00 per issue in the
U.S. $70.00 per year or $14.00 per issue else-
where. Back volumes are available at
$78.00 per volume ($13.00 per issue) in the
U.S. Add $5.00 more per volume of $1.00
per issue elsewhere.

Crop Science
The Crop Science Society of America's bi-
monthly journal covers all aspects of crop
development in detail. Original research
findings in ecology; cytology; statistics;
weed science; crop^quality; crop breeding
and genetics; crop physiology and bio-
chemistry; crop and seed production; and
crop and crop growth utilization are pre-
sented
Letters, invited reviews and notes on
equipment, experimental techniques, and
observations are also presented in this
recognized journal.
$65.00 per ye*ar or $13.00 per issue in the
U.S. $70.00 per year or $14.00 per issue else-
where. Back volumes are available at
$78.00 per volume ($13.00 per issue) in the
U.S. Add $5.00 more per volume or $1.00
per issue elsewhere.

AMERICAN SOCIETY OF AGRONOMY CROP SCIENCE SOCIETY OF AMERICA

677 South Segoe Road, Madison, Wisconsin 53711
 NOTE TO CONTRIBUTORS
All articles must comply with the following
requirements:
Manuscripts
Papers should be written as concisely as pos-
sible and should be typed in double-spacing on
one side of white quarto (A4 size - 210x297
mm) and with a margin of 4 cm on all sides.
The original typescript and two carbon copies
should be submitted to the Chief Editor of
PERTANIKA. Complicated mathematical ex-
pressions or chemical formula should be care-
fully typed and thoroughly checked. Tables and
figures should appear on separate sheets of
paper. Articles should not exceed 10 printed
pages - ie about 30 quarto pages of typewritten
scripts. Space for figures and tables must be
estimated separately. They are reduced to about
half their size on the printed page.
Communication
A Communication is intended for the rapid
publication of timely and significant work. The
manuscript should not exeed five double -
spaced typewritten pages including any figures
and tables but excluding title, abstracts and
references. It should contain sufficient experi-
mental detail to allow repetition of the work
and should include a brief statement on the
significance of the work.
Abstract
Abstracts in Bahasa Malaysia and English, each
not exceeding 200 words, should follow imme-
diately after the title of the paper. Papers from
outside of Malaysia may be submitted with an
English abstract only.
References
References are to be cited in the text with the
year of publication, arranged in alphabetical
order and listed at the end of the text giving
the author's name and initials followed by the
year of publication, title of the paper, title of
e periodical, volume and page.
e.g TIMMS, R.E. 1978. Artefact Peaks in the
Preparation and Gas-liquid Chromatographic
Determination of Methyl Esters. Aust Dairy
TechnoL 33 (1): 4-6. Abbreviations of titles of
periodicals should conform to those used in "A
World List of Scientific Periodicals".
Tables
Tables should be numbered with Arabic
numerals and each table must be referred to in
the text. Titles and column headings should be
brief, and descriptive matter in the tables
confined to a minimum.
Footnotes
Footnotes are to be numbered consecutively
and typed at the bottom of the page below a
horizontal rule.
In tables, references to headings and figures
should be indicated by superscripts in alpha-
betical order. Explanatory notes should be typed
immediately below the table.
Illustrations
a. All figures should be numbered consecu-
tively with Arabic numerals, and each figure
should be referred to in the text. The title of
the paper and the figure number should be
written on the back of the original copy in
pencil or (preferably) on a gummed label.
Figures should be about twice the size of the
printed reproduction.
b. Photographs: All photographs (glossy back-
and-white prints) should be supplied with ap-
propriate scales.
c. Lettering and numbering on the face of the
figure itself should be kept to a minimun.
Unit of measure
Metric units must be used for all measurements.
Reprints
Authors will receive 20 copies of each article
free.