Journal of

Agricultural Science
and Technology B
Volume 2, Number 3, March 2012 (Serial Number 11)

David Publishing

David Publishing Company

www.davidpublishing.com
Publication Information:
Journal of Agricultural Science and Technology B (Earlier title: Journal of Agricultural Science and Technology, ISSN
1939-1250) is published monthly in hard copy (ISSN 2161-6264) by David Publishing Company located at 9460 Telstar
Ave Suite 5, EL Monte, CA 91731, USA.

Aims and Scope:

Journal of Agricultural Science and Technology B, a monthly professional academic journal, particularly emphasizes new
research results in agronomy, forestry, aquaculture, fisheries, food science, all aspects of crop physiology, modeling of crop
and forest systems, engineering solutions and so on. Articles interpreting practical application of up-to-date technology are
also welcome.

Editorial Board Members (in alphabetical order):

Catherine W. Gitau (Australia) Chuah Tse Seng (Malaysia) Dharmatilleke Bandula Kelaniyangoda (Sri Lanka)
Ekachai Chukeatirote (Thailand) Erin K. Espeland (USA) Farzana Perveen (Pakistan)
Francesco Montemurro (Italy) Gulshan Mahajan (India) Idress Hamad Attitalla (Libya)
Jang Ho Son (Korea) Jagadish Timsina (Bangladesh) Jelena Bokovi (Serbia)
Manoj K. Shukla (USA) Mehmet Musa zcan (Turkey) M. S. Qureshi (Pakistan)
Milad Manafi (Iran) Mehmet Rt Karaman (Turkey) Noureddine Benkeblia (Algeria)
Natraj Krishnan (USA) Olivier A. E. Sparagano (France) Renato S. Pacaldo (USA)
Ram C. Bhujel (Thailand) Shoil M. Greenberg (USA) Sanjeev Kumar Chauhan (India)
Shri Mohan Jain (Finland) Thai Ngoc Chien (Vietnam) T. Chatzistathis (Greece)
Vasudeo P. Zambare (USA) Vasileios Fotopoulos (Greece) Young Jung Kim (Korea)
Yusuf Bozkurt (Turkey) Zeki Candan (Turkey)
Manuscripts and correspondence are invited for publication. You can submit your papers via web submission, or E-mail to
agriculture@davidpublishing.org. Submission guidelines and web submission system are available at
http://www.davidpublishing.org, http://www.davidpublishing.com.

Editorial Office:
9460 Telstar Ave Suite 5, EL Monte, CA 91731, USA
Tel: 1-323-984-7526
Fax: 1-323-984-7374
E-mail: agriculture@davidpublishing.org, agriculture66@hotmail.com, agriculture@davidpublishing.com
Copyright2012 by David Publishing Company and individual contributors. All rights reserved. David Publishing
Company holds the exclusive copyright of all the contents of this journal. In accordance with the international convention,
no part of this journal may be reproduced or transmitted by any media or publishing organs (including various websites)
without the written permission of the copyright holder. Otherwise, any conduct would be considered as the violation of the
copyright. The contents of this journal are available for any citation. However, all the citations should be clearly indicated
with the title of this journal, serial number and the name of the author.

Abstracted/Indexed in:
Database of EBSCO, Massachusetts, USA
Chemical Abstracts Service (CAS), USA
Cambridge Scientific Abstracts (CSA), ProQuest Science Journals, USA
Ulrichs Periodicals Directory, USA
Summon Serials Solutions, USA
Chinese Database of CEPS, American Federal Computer Library Center (OCLC), USA
Chinese Scientific Journals Database, VIP Corporation, Chongqing, China

Subscription Information:
Price (per year)
Print $1200, Online $960
Print and Online $1600
David Publishing Company
9460 Telstar Ave Suite 5, EL Monte, CA 91731, USA
Tel: 1-323-984-7526 Fax: 1-323-984-7374
E-mail: order@davidpublishing.org

D DAVID PUBLISHING

David Publishing Company

www.davidpublishing.com
 Journal of
Agricultural Science
and Technology B
Volume 2, Number 3, March 2012 (Serial Number 11)

Contents
Research Papers

287 Techno-economic Appraisal of Biodiesel from Jatropha Curcas: An Egyptian Case Study
Shadia R. Tewfik, Salwa I. Hawash, Nahed K. Atteya, Guzine I. El Diwani and Ihab H. Farag
298 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability of Commercial
Wheat, Rice or Waxy Rice Flour
Yuthana Phimolsiripol
307 Use of High Content Fish Silage Wet Food in Final Growth out of Pac (Piaractus mesopotamicus,
Holmberg 1887) in Northeast Argentina
Gustavo Wicki, Oscar Galli Merino, Pablo Cal and Facundo Sal
312 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics under
Different Forest Management in the Sierra de Francia, Salamanca, Spain
Sergio Salazar, Luz-Estela Snchez, Purificacin Galindo and Ignacio Santa-Regina
329 Linking Markets to Smallholder Agro-forestry Farmers as a Strategy for Poverty Alleviation in
the Tropics
Twaha Ali Basamba, Barnabas Kiiza, Clement Mayanja, Bob Nakileza, Frank Matsiko, Paul Nyende
and Elizabeth Bacwayo Kukunda
339 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The Relationship between
Silvicultural Activity and Quality of Stand
Chaw Chaw Sein, Hoang Lien Son and Ralph Mitlhner
349 Assessment of Grain Damage and Weight Loss on Farm Stored Maize in Highlands Areas of
Bungoma District, Kenya
Kimondo Mutambuki and Christopher Mugo Ngatia
362 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR
Markers
Lucky O. Omoigui, Mary Yeye, Boukar Ousmane, Bhavani S. Gowda and Michael P. Timko
368 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield in Uganda: A Propensity
Score Matching Approach
Margaret Nabasirye, Barnabas Kiiza and George Omiat
378 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate Composition of
Caspian Sea White Fish (Rutilus frisii kutum) Fillets
Sahar Jalili
385 The Influence of Low Frequency Electromagnetic Field (LF EMF) on the Agricultural Crops
Seeds Germination
Mihail Gennadevich Barishev, Stepan Sergeevich Dzhimak, Gennadij Ivanovich Kasjanov and Denis
Igorevich Sashhkov
Journal of Agricultural Science and Technology B 2 (2012) 287-297
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Techno-economic Appraisal of Biodiesel from Jatropha

curcas: An Egyptian Case Study

Shadia R. Tewfik1, Salwa I. Hawash1, Nahed K. Atteya1, Guzine I. El Diwani1 and Ihab H. Farag2
1. Department of Chemical Engineering and Pilot Plant, National Research Center, El-Tahrir St. Dokki, Cairo, Egypt
2. Department of Chemical Engineering, University of New Hampshire, Durham, NH, USA

Received: September 15, 2011 / Published: March 20, 2012.

Abstract: Biodiesel (fatty acid methyl ester) has been identified as a non-toxic biodegradable alternative fuel that is obtained from
renewable sources. Over the last decade, there has been increasing interest in producing biodiesel from conventional sources such as
soybean, canola, sunflower, and coconut oils. Current efforts are directed towards the development of new non-edible resources.
Among these Jatropha Curcas comes at the forefront. In Egypt, Jatropha Curcas has grown successfully using primary treated
wastewater. Also, extensive R&D efforts identified the optimum conditions for the various processing stages namely crushing,
extraction, transesterification and purification. Based on the research findings, the techno-economic appraisal of biodiesel production
from Jatropha Curcas is conducted. Two nominal capacities namely 8,000 and 50,000 metric tons/yr have been proposed. Several
scenarios have been formulated to take into consideration varying productivity (3.4 to 5.8 ton fruits per 4,000 m2 (acre)) and varying
recovery rates of oil from seeds. Economic indicators including capital and production costs for the various processing stages and
revenues according to current prices of oil and cake have been obtained. The price of biodiesel that provides a simple rate of return
(SRR) on investments of 10% was in the range of $0.3-0.7/liter for the different assumed scenarios which is lower than the prevailing
price of biodiesel (about $1/liter) in the US. Thus, in view of experimental results and economic assumptions, there are positive
prospects for the production of biodiesel from Jatropha Curcas under Egyptian conditions.

Key words: Biodiesel, Jatropha curcas, experimental, simulation, economic indicators.

1. Introduction Biodiesel as a fuel runs in any unmodified diesel

engine and it has been shown to give engine
Biodiesel is a liquid fuel made from any vegetable
performance generally comparable to that of
oil including oils pressed straight from the seeds
conventional diesel fuel while reducing engine
(virgin oils) such as soybean, sunflower, canola,
emissions of particulates, hydrocarbons and carbon
coconut and Jatropha Curcas (JC) [1-3]. It can be
monoxide [6]. It has been reported that when biodiesel
produced from any material that contains fatty acids,
was used, CO emissions was reduced by 8.6%,
whether it is linked to other molecules or present as
hydrocarbons by 30.7%, and particulate emissions by
free fatty acids. Thus, various vegetable fats and oils,
63.3% [1]. However, CO2 emissions were increased
animal fats, waste greases, and cooking oil [4] can be
by 2.6% and NOx emissions were increased by 5% [7].
used as feedstock for biodiesel production [5]. The
The technology to produce methyl ester for use as
choice of feedstock is based on several variables such
fuel has been commercially used since the early 1990s
as local availability, cost, government support and
[8]. Biodiesel is typically mixed with petroleum diesel
performance as a fuel.
in blends between 5% and 50% (designated by a B
followed by the percentage of biodiesel in the blend).
Corresponding author: Guzine I. El Diwani, professor,
Ph.D., research fields: chemical and environmental engineering, Any biodiesel blend (even B100) can be used in
polymer engineering, biodiesel chemistry and engineering. standard diesel engines without engine modification [5].
E-mail: geldiwani@yahoo.com.
288 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study
There are several benefits for using biodiesel as a
blended fuel in diesel engines [9, 10], such that most
major car and equipment manufacturers have issued
statements certifying that blends up to B20 are
acceptable for use in their diesel engines without loss of
warranty [10]. The economics of biodiesel production
from several oil seeds in general and from JC in
particular have been published in several reports issued
worldwide [11, 12]. This work is concerned with the
techno-economic appraisal of biodiesel production
from JC. Although several articles and reports have
addressed the issue, appraisal under Egyptian
conditions is highly needed for application in the
country, in addition to providing guidelines for
adoption under similar conditions worldwide. In this
paper, the first section includes the technical aspects
obtained through an extensive R&D program. It is then
followed by simulation and basic engineering of the
crushing, extraction, transesterification and purification
stages. The third section includes the financial aspects
and analysis through the formulation and assessment of
several scenarios taking into consideration the
productivity of seeds per unit area and the extent of
recovery of raw oil from the JC seeds.
2. Basis of Technical Aspects
2.1 Jatropha Curcas Plantation in Egypt
Jatropha plant is cultivated in different places in
Upper (Sothern) Egypt and in Egyptian deserts.
Jatropha trees were irrigated with primary treated
municipal wastewater. Average productivity of
Jatropha fruits was 3.7 metric tons fruit/4,000 m2 (acre)
with average oil recovery of 25% by weight from seeds.
2.2 Optimum Biodiesel Production Conditions for
Jatropha Plant
Through extensive experimental work on bench and
pilot scale, optimum operating conditions for Jatropha
oil and biodiesel production have been obtained [13]
and are taken as the basis for this work as outlined
below.
2.2.1 Jatropha Oil Extraction
The extraction process essentially comprises fruit
dehulling, seed crushing to 2 mm, extraction of oil
using hexane as solvent at ratio 1:5 seeds to hexane
[14, 15] and separation of hexane from oil by hexane
evaporation under vacuum at 40 C.
2.2.2 Transesterification
Conditions for transesterification include:
(1) Use of methanol as alcohol for transesterification
at molar ratio of 6 to 1, alcohol to oil;
(2) Use of sodium hydroxide as catalyst at a 0.75%
or potassium hydroxide at 1% based on oil weight;
(3) Mix the reaction mixture at 60 rpm for 10
minutes and continue mixing for one hour at a
temperature 65 C to complete the reaction;
(4) 94%-98% reactant oil conversion to biodiesel is
achieved and the overall reaction product is left to
settle in a clarifier for two hours;
(5) Glycerol is obtained as byproduct at a rate of
about 10% of produced biodiesel and settled as the
lower phase in the reaction product at 50% purity;
(6) This lower phase is separated and purified using
phosphoric acid to obtain glycerol of 85% purity;
(7) Biodiesel and unreacted methanol constitute the
upper layer in the clarifier;
(8) Methanol which has to be separated using a
distillation column is to be recycled;
(9) Biodiesel is washed with water for the
elimination of any soap formed, and tested;
(10) The material balance based on pilot plant
experiments is depicted in Fig. 1.
2.2.3 Seed Cake Evaluation
(1) Seed cake obtained from oil extraction which
constitutes about 55%-75% of the seed, is dried at
60 C to eliminate hexane traces; positive results for
use as fertilizer have been obtained.
(2) Dried seed cake is also tested as animal fodder
and proved successful when substituting conventional
rabbit fodder by 7.5%. It is to be noted that the content
of phorbol ester of 0.085%, which is the cause of
toxicity, can be removed by special treatment [16].
 Techno-economic Appraisal of Biodiesel from Jatropha curcas: 289
An Egyptian Case Study

Fig. 1 Material balance based on pilot-scale experimental results.

Further investigations are to be undertaken for
treating the seed cake and testing for use as animal
fodder.
2.2.4 Treatment and Recovery of Side-Streams
It is very important to recover and recycle the
excess methanol from both main product (biodiesel)
and byproduct (glycerol). Also glycerol is a very
valuable byproduct in this process, so it is useful to
refine for use in many industries specifically for
medical purposes.
The basic processing steps for biodiesel production
from 4,000 m2 (one acre) Jatropha plantation are
shown in Fig. 1 noting that the mass balance is based
on the actual experimental results. Various plausible
options are investigated in the following sections.
3. Techno-economic Aspects for Biodiesel
Production under Egyptian Conditions
3.1 Capacity of Production Facilities
Biodiesel is produced commercially in large
290 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study

capacities in US and Europe. Plants up to 85 million developed a process model to estimate biodiesel
gallons/year (about 275 thousand metric tons/year) production costs. Simulation has been conducted
production are in operation in the US [17]. Conversely, using Aspenplus from Aspen Tech. The software from
in developing countries, biodiesel production is still at Haas has been used as a basis after introducing
small scale level. However, the potential for necessary modifications to adopt for the raw materials,
production in large scale facilities is high with the process conditions and capacities proposed in this
increasing interest in JC and other plants or crops work. The simulation includes the following section:
grown in South Asia and Africa. This is justified by seed crushing, oil extraction, transesterification,
the high productivity of Jatropha and other plants in biodiesel purification and glycerol purification. The
these areas. revised process flowsheet is depicted in Fig. 2.
Thus, within the scope of this work an economic Simulation output including material and energy
model and indicators for two nominal annual balance and specifications of essential components
capacities namely 8,000 and 50,000 metric tons have marked on Fig. 2 is outlined in Table 1 for the 50,000
been developed. The optimum process conditions, as tons/yr production capacity. Results have been used
developed in our laboratory, outlined in section 2 are for the cost estimation of capital costs for the two
adopted. proposed capacities respectively.

3.2 Simulation and Basic Engineering 3.3 Basis of Estimates for the Economic Model

Simulation and basic engineering of Although several reports on the economics of

extraction/transesterification of raw oil for the two biodiesel production on the international level have
proposed capacities have been conducted. Previous been published [11, 12], the financial aspects relevant
work on simulation of biodiesel has been reported to production of biodiesel under Egyptian conditions
using Hysys from Hyprotech for four proposed need to be assessed. The basis of estimates for
technologies using cooking oil [18]. Haas et al. [19] proposed two options are outlined below.
MIXER1
EXTRACT MEOHREC
MEOHWAT
HXANE
SEPAR MEOH
HEXANE- MIX1
PUMP1 MIX3 WATER WASHCOL FAME
NAOH MEOHCOL
CRUSHER MIX2
CAKE REACOUT PUMP3 EST4
MIXER2 ESTCOL
B3
OIL
CRASHED RECY-HEX OIL3
MIXER PUMP2 REACTOR EST2
HX1 EST1 2 EST3
HEXANE
OIL2
AQU1 WATMEOH

HX2
NEUTR
GLYCRCOL

AQU3
H3PO4 GLYCEROL
AQU2
1

FILTER

SOLIDS
B2
OILREC

Fig. 2 Flow sheet of the complete crushing/extraction/transesterification process.

 Techno-economic Appraisal of Biodiesel from Jatropha curcas: 291
An Egyptian Case Study

Table 1 Material & energy balance and specifications of essential components for 50,000 metric ton per year.
a.Methanol Recovery Column:- b. Biodiesel purification column: Component split fraction

Component Split Fractions Outlet Stream

Outlet streams MEOHREC EST1 Meohwat Biodiesel Oilrec
Methanol 0.94 6.00E-02 methanol 0.9371 6.29E-02 2.82E-14
Oil 3.52E-03 0.99648 oil 1.43E-05 3.59E-03 0.9964
Biodiesel 1.83E-15 1 biodiesel 6.38E-04 0.97293 2.64E-02
Glycerol 1.89E-16 1 glycerol 0.11152 0.85827 3.02E-02
NaOH 0.11422 0.88578 naoh 0.93417 6.58E-02 9.51E-16
Water 0.11422 0.88578 water 0.93417 6.58E-02 9.51E-16
Summary of key results Summary of key results

Number of stages 7 Number of 8

theoretical stages

Top stage 29.2638 top stage oC 154.101

o
temperature C temperature
Bottom stage 65.433 bottom stage oC 287.394
o
temperature C temperature
Top stage liquid kmol/hr 73.5179 top stage liquid kmol/hr 23.6998
flow flow
Bottom stage liquid kmol/hr 36.8265 bottom stage liquid kmol/hr 0.79661
flow flow
Bottom stage vapor kmol/hr 56.9743 top stage vapor kmol/hr 1.39521
flow flow
Molar reflux ratio 2 bottom stage vapor kmol/hr 104.632
flow
Molar boilup ratio 1.5471 molar reflux ratio 0.94817

Condenser duty (w/o kw -775.115 molar boilup ratio 131.346

subcool)
Reboiler duty kw 768.907 condenser duty kw -2,055.33
(w/o subcool)
reboiler duty kw 2,519.85
Oil feed pump: c. Glycerol column:
Input data: Component split fractions outlet streams

Outlet pressure bar 4 watmeoh glycerol

Driver efficiency 1 methanol 1 4.55E-06

Results biodiesel 0.36749 0.63251

Volumetric flow rate 8.31471 glycerol 0.017395 0.9826

cum/hr
Pressure change bar 3.8 water 1 1.55E-09

Npsh available 2.33736 Summary of key results

meter
Fluid power kw 0.87766 top stage oC 68.9594
temperature
Brake power kw 2.23315 bottom stage oC 269.036
temperature
Electricity kw 2.23315 top stage liquid kmol/hr 65.8132
flow
Pump efficiency 0.39302 bottom stage liquid kmol/hr 7.68049
used flow
Net work required 2.23315 bottom stage vapor kmol/hr 76.0203
kw flow
Head developed 44.4806 molar reflux ratio 2
meter

Option (I) essential processing stages namely plantation,

In this option, it is assumed that each of the crushing & extraction, and transesterification &
292 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study
purification could be considered as a sub-project and
financed separately.
Option (II)
In this option II, it is assumed that the whole project
is integrated, i.e., seeds from the plantation stage are
delivered directly to the extraction stage with no profit.
Similarly, the raw oil is delivered directly to the
transesterification stage.
All other assumptions are similar to option I except
for some consequent assumptions such as the decrease
in working capital and operating costs corresponding
to no direct profits for seeds or raw oil.
The basic assumptions for both cases are presented
below:
(1) Plantation
Productivity: two alternatives
a. High productivity obtained at demonstration areas
b. Average reported productivity on international
level
(2) Crushing/Extraction/Transesterification
Seeds in fruit: 58%
Oil recovery (weight percent of raw oil extracted
from seeds): three options: 25%, 35%, 45%.
(3) Oil recovery at high productivity
a. High recovery: 45% of seeds/high productivity
Tons raw oil per 4,000 m2 (acre) 1.46
Tons biodiesel per 4,000 m2 (acre) 1.38
Tons cake per 4,000 m2 (acre) 1.78
b. Low recovery: 25% of seeds/high productivity
Tons raw oil per 4,000 m2 (acre) 0.81
Tons biodiesel per 4,000 m2 (acre) 0.76
Tons cake per 4,000 m2 (acre) 2.43
Corresponding values for average productivity and
average recovery have been estimated.
(4) Raw materials-intermediate products and
products/prices
The prices of raw materials, intermediate products
and products according to current market prices for
raw materials and indicated basis of estimates for
intermediate products and products are presented on
Table 3.
Table 2 Alternative assumptions for high and average
productivities of fruits and seeds.
Average High
Item
productivity productivity
3.43 5.58 Ton fruits per 4,000 m2 (acre)
2.0 3.2 Tons seed per 4,000 m2 (acre)
(5) Capital investments
The capital investments for the process components
have been estimated as follows:
a. Plantation
Investment costs according to actual
implementation in Luxor in the South of Egypt as
provided by the Ministry of Agriculture which is the
implementing agency have been adopted.
b. Crushing and extraction
Economic data published for the crushing and
extraction stage [11] have been adopted after
modifications to take into consideration, among others,
the local prices of some components, costs incurred
for transportation, etc.
The working capital has been assumed to be 25% of
annual operating costs.
c. Transesterification and purification
The prices of equipment have been estimated from
several reported sources [20-22]. Prices have been
adjusted to the year 2010 by cost indices. Other
components of fixed capital have been estimated as
presented in Table 4.
(6) Operating costs
Raw materials costs have been estimated according
to the actual consumption from pilot experimental
results and prevailing costs. Other components of
operating costs are presented in Table 5.
(7) Depreciation
Depreciation has been estimated using straight line
method and assuming a lifetime of 15 years for the
facilities. Two scenarios have been assumed for each
option as depicted in Fig. 3.
(8) Profitability
The profitability is expressed as the simple rate of
return (SRR) for a typical year (year 8) which is
defined by the following equation:
 Techno-economic Appraisal of Biodiesel from Jatropha curcas: 293
An Egyptian Case Study

Table 3 Estimated prices of raw materials and products.

Basis of estimates Price $/ton Item
Market price 791 Hexane
Market price 818 Methanol
Market price 455 NaOH
Market price 729 H3PO4
Actual production costs *1.25 to account for profits for the plantation stage) 145-230 Seeds
Prices that ensure a Simple Rate of Return on Investment (SRR) of 12%-15% for the
300-625 Raw oil
crushing/extraction stage
US price (November 2010) 945 Biodiesel
Market price of equivalent product used as animal fodder 145 Cake
Market price 545 Glycerol

I
Table 4 Basis of estimates of capital costs. High Productivity
II
Average Productivity
(5.58 ton fruit/acre)
Item % Relative to High Recovery
(3.43 ton fruit/acre)
High Recovery
(45% oil )
Fixed capital (45% oil )

Purchased equipment 100 Purchased equipment

Equipment setting 30 Purchased equipment
Piping 15 Purchased equipment IV III
Average Productivity High Productivity
Civil 15 Purchased equipment (3.43ton fruit/acre) (5.58 ton fruit/acre)
Steel 15 Purchased equipment Average Recovery Low Recovery
(35% oil ) (25% oil )
Instrumentation 10 Purchased equipment
Electrical 10 Purchased equipment Fig. 3 Scenarios for the two capacities: 8,000 and 50,000
Insulation 8 Purchased equipment tons/yr.
Paint 8 Purchased equipment
Other 10 Purchased equipment
Engineering 10 Total fixed capital
Contract fee 5 Total fixed capital
Contingencies 10 Total fixed capital
Working capital 25 Annual operating costs

Table 5 Basis of estimates of operating costs others than

raw materials.
Labour $/Annum
Engineers 4,364 Fig. 4 Area requirements (acres, 1 acre = 4,000 m2).
Supervisors 3,273
Administration 2,182
Labourer 1,091
Maintenance 2% Of capital
Others 5% Of total annual operating cost

SRR = (Annual revenues-annual operating

costs-depreciation)/Total investments 100

4. Results and Analysis

4.1 Area Requirements

The area requirements for the assumed scenarios

Fig. 5: Distribution of annual operating cost for
are presented in Fig. 4. SClll-50,000 ton/yr.
294 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study

4.2 Financial Indicators for Option I
4.2.1 Annual Operating Costs
The distribution of annual operating costs over its
components for the transesterification stage is depicted
in Fig. 5 for SCIII for the 50,000 tons/yr capacity.
Rather similar patterns are valid for other scenarios
and for the smaller capacity. It is apparent that the raw
materials constitute the major component of the
annual operating costs.
40
4.2.2 Estimated Investments, Production Costs,
Profits and SRR
The estimated total capital investment, annual
operating costs, total profits and hence the SRR on
investments are presented in Figs. 6 and 7 for the four
assumed scenarios for the 8,000 and 50,000 tons
biodiesel/annum respectively. For the 8,000
ton/annum, the investments varied between $14 and
$24 million for the different scenarios respectively.
Corresponding values for the 50,000 ton/annum varied

35

30
T ot a l C a pi t a l
I nv e s t me nt s M i l l i ons
$
25
T ot a l P r oduc t i on
C os t s M i l l i ons $

20
T ot a l P r of i t s
M i l l i ons $

15
A v e r a ge SR R %

10

0
SC I SC II SC III SC IV

Fig. 6 Financial indictors for 8,000 ton/yr biodiesel.

160

140

120
Total Capital Investments
Millions $
100
Total Production Costs
Millions LE
80 Total Profits Millions$

60 Average SRR%

40

20

0
SC I SC II SC III SC IV
Fig. 7 Financial indicators for 50,000 ton/year biodiesel.
 Techno-economic Appraisal of Biodiesel from Jatropha curcas: 295
An Egyptian Case Study

between $76 and 138 million respectively. The total
profits varied between $5 & 7 for the 8,000
tons/annum and $30 & 45 for the 50,000 tons/annum.
The SRR varied between 19% & 36%, and 22% &
44% for 8,000 and 50,000 tons/annum, respectively.
4.2.3 Distribution of Capital Costs over Processing
Stages
The distribution of capital costs for the two
capacities is demonstrated in Figs. 8 and 9.
Plantation represents the major cost component.
4.2.4 SRR% for Transesterification at Prevailing
Biodiesel Price
For a price of biodiesel of $945/ton, which is
currently prevailing [23], the SRR is presented in Fig.
10. It is apparent that for the transesterification stage,
the SRR% varies between 80% and 176%.
4.2.5 Prices of Biodiesel for a Value of SRR of 10%
Since the national prevailing discount factor is 10%,
the price of biodiesel that provides SRR of 10% has
been estimated as presented in Table 6.
The estimated prices for a SRR that is equivalent to
prevailing discount factor (interest rate) are below the
current prevailing international prices of biodiesel.
4.3 Financial Indicators for Option II
In option II, it is assumed that the whole project is
integrated, i.e., seeds from the plantation stage are
delivered directly to the extraction stage with no profit.
Similarly, the extracted raw oil is delivered directly to
the transesterification stage.

Fig. 8 Distribution of capital costs over processing

stages% for 8,000 ton/yr capacity.

Fig. 10 SRR% for transesterification at constant biodiesel

price $945/ton.

Table 6 Specific biodiesel price for transesterification

(SRR 10%).
SC IV SC III SC II SC I
Capacity
$/ton
781 575 674 459 8,000 ton/yr
746 541 642 425 50,000 ton/yr
$/liter Capacity
Fig. 9 Distribution of capital costs over processing 0.70 0.52 0.61 0.41 8,000 ton/yr
stages% for 50,000 ton/yr capacity. 0.67 0.49 0.58 0.38 50,000 ton/yr
296 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study

All other assumptions are similar to option I except Table 7 Economic indicators for integrated scheme-option
II.
for some consequent assumptions such as decrease in
Annual biodiesel production 8,000 Tons
working capital and operating costs corresponding to
Ltem/Scenario SCI SCII SCIII SCIV
no direct profits for seeds or raw oil. Total capital investments $ 11.9 17.6 18.1 21.1
Results are summarized in Table 7. SRR% is given Total production Costs $ 1.2 1.2 5.5 5.9
for a constant estimated price of LE $945/ton Total profits $ 5.3 4.2 6.3 4.9
Average SRR% 44.5 24.0 35.1 23.5
biodiesel while in Table 8 the specific biodiesel prices
Annual biodiesel production 50,000 Tons
for SRR = 10% are given. As anticipated, relatively
Ltem/Scenario SCI SCII SCIII SCIV
lower prices for biodiesel could give a SRR of 10% Total capital investments millions 65.1 65.1 97.4 119.7
for option II as no profits for agriculture and Total Production costs millions $ 6.7 6.7 33.9 36.5
extraction stages have been assumed. Total profits millions $ 34.0 27.7 39.8 31.2
Average SRR% 52.2 28.4 23.5 26.1
5. Conclusions
Tabie 8 Prices of biodiesel for SRR 10%-integrated
With the success of plantation of Jatropha Curcas system-option ll.

under Egyptian conditions using primary treated SCIV SCIII SC II SC I

Capacity
municipal wastewater, an extensive experimental $/ton
610 380 656 435 8,000 ton/yr
program for crushing, extraction and
584 357 607 396 50,000 ton/yr
transesterification defined the optimal and suboptimal
$/liter Capacity
conditions for each processing stage. Accordingly,
0.7 0.52 0.61 0.41 8,000 ton/yr
simulation and basic engineering have been conducted
0.67 0.49 0.58 0.38 50,000 ton/yr
to determine the requirements for assumed two
capacities namely 8,000 and 50,000 metric tons 50,000 tons/annum, respectively. Positive economic
biodiesel/annum. In view of the wide variation of indicators have also been obtained if it is assumed that
reported and actual figures for productivity of fruits, all stages are considered as an integrated project.
and extracted raw oil from seeds, several scenarios The price of biodiesel that provides a SRR of 10%
have been assumed to take into consideration the was in the range of $0.3-0.7/liter for the different
possible combinations of high (5.58 ton fruits/4,000 assumed scenarios which is lower than the prevailing
m2 (acre)) and average (5.58 ton fruits/4,000 m2 (acre)) price of biodiesel (about $1/liter)
productivity with 25, 35, 45% recovery respectively. Thus, in view of experimental results and economic
When it is assumed that each of the essential assumptions, there are positive prospects for
processing stages namely plantation, crushing & production of biodiesel from Jatropha Curcas under
extraction, and transesterification and purification Egyptian conditions.
could be considered as a sub-project and financed
Acknowledgment
separately, for the 8,000 tons/annum, the investments
varied between $14 and $24 million for the different This work has been undertaken by the National
scenarios, respectively. Corresponding values for the Research Center, Engineering Research Division,
50,000 tons/annum varied between $76 and $138 Chemical Engineering and Pilot Plant Department.
million, respectively. The total profits varied between The authors acknowledge the US-STDF Grant, USDA
$5 & 7 million for the 8,000 ton/annum and $30 & 45 Grant 58-3148-8-174 and University of New
million for the 50,000 tons/annum. The SRR varied Hampshire for its support. The authors have declared
between 19% and 36%, and 22% and 44% for 8,000 and no conflict of interest.
 Techno-economic Appraisal of Biodiesel from Jatropha curcas:
An Egyptian Case Study
References
[1] A. Demirbas, Biodiesel from vegetable oils via
transesterification in supercritical methanol, Energy
Conversion and Management 43 (2002) 2349-2356.
[2] N. Foidl, G. Foidl, M. Sancher, M. Mittebach, S. Hockel,
Jatropha curcas as a source for the production of biofuel
in Nicaragua, Bioresource Technology 58 (1996) 77-82.
[3] S.K. Datta, R.K. Pandey, Jatropha curcasA promising
crop for new source of fuel, Applied Botany 17 (1993)
108-118.
[4] A.A. Refaat, N.K. Attia, H.A. Sibak, S.T. Sheltawy, G.El.
Diwani, Production optimization and quality assessment
of biodiesel from waste vegetable oil, Int. J. Environ. Sci.
Tech. 5 (1) (2008) 75-82.
[5] G. Mangesh Kulkarni, Ajay K. Dalai, Waste cooking
oilAn economical source for biodiesel: A review, 1nd.
Eng. Chem. Res. 45 (2006) 2901-2913.
[6] M.S. Graboski, R.L. McCormick, Combustion of fat and
vegetable oil derived fuels in diesel engines, Prog.
Energy Combust. Sci. 24 (1998) 125-164.
[7] Y. Ulusoy, Y. Tekin, The engine tests of biodiesel from
used oil, Energy Sources 26 (2004) 927-932.
[8] R. Anthony, Biodiesel, Performance, Costs, and Uses,
Energy Information Administration, 2004.
[9] S. Lele, Biodiesel, Biodiesel 10 TPD, Report, 2005, pp. 1-13.
[10] J. Deere, Biodiesel in John Deere Tractrors,
http://www.deere.com/en_US/ag/servicesupport/tips/tract
ors/9000/Biodiesel in John_ Deere_ Tractors html.
[11] Economics: Jatropha Biodiesel,
http://www.jatrophaworld.org/15.html.
[12] Biodiesel in British Columbia, Feasibility Study Report,
Prepared for Eco-Literacy Canada, in: Schaddelee (Ed.),
297
WISE Energy Co-op, Mike Boyd, Anita Murray-Hill,
May, 2004.
[13] G.El. Diwani, S. Hawash, N. Kamal, Development and
Evaluation of biodiesel fuel and byproducts from
Jatropha oil, Int. Journal of Environmental Science and
Technology 6 (2) (2009) 219-224.
[14] N. Kaushik, S. Kumar, Jatropha curcas L. Silviculture
and Uses, Dr. Updesh Purohit for Agrobios, India, 2006.
[15] N. Jones, J.N. Miller, Jatropha curcasA multipurpose
species for problematic sites, Land Resources Series Asia
Technical Dept, World Bank 1 (1992) 1-40.
[16] G.I.El. Diwani, Sh.A.El. Rafei, S.I. Hawash, Ozone for
phorbol esters removal from Egyptian Jatropha oil seed
Cake, Advances in Applied Science Research 2 (2) (2011)
221-232.
[17] Biodiesel Plants, http://www.card.iastate.edu/
research/bio/tools/biodiesel.aspx.
[18] Y. Zhang, M.A. Dub, D.D. McLean, M. Kates,
Biodiesel production from waste cooking oil: 1. Process
design and technology assessment, Bioresource
Technology 89 (2003) 1-16.
[19] H.J. Michael, A.J. McAloon, W.C. Yee, T.A. Foglia, A
process model to estimate biodiesel production costs,
Biosource Technology 97 (2006) 671-678.
[20] The Economics & Sustainability of Bio-diesel,
http://www.bdpedia.com/biodiesel/economics/economics.
html.
[21] M.S. Peters, T.D. Klaus, R.E. West, Plant Design and
Economics for Chemical Engineers, Mcgraw-Hill, 2002.
[22] Aspen Icarus, ASPENTECH [Online], 2006,
http://www.aspentech.com.
[23] http://www.icispricing.com/il_shared/Samples/SubPage1
0100131.asp.
Journal of Agricultural Science and Technology B 2 (2012) 298-306
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Effect of Malva Nut Gum on Pasting Characteristics and

Freeze-thaw Stability of Commercial Wheat, Rice or
Waxy Rice Flour

Yuthana Phimolsiripol
Division of Product Development Technology, Faculty of Agro-Industry, Chiang Mai University, Chiang Mai 50100, Thailand

Received: September 19, 2011 / Published: March 20, 2012.

Abstract: The objective of this research was to investigate the effect of Malva nut gum (MG) replacement on the pasting
characteristics and freeze-thaw stability of wheat, rice or waxy rice flours. Pasting properties and freeze-thaw stability of different
flours incorporated with 0, 0.5%, 1%, 2%, 3% and 5% of MG were investigated. Pasting temperature (60 C-87 C) of the pastes
significantly decreased with increasing of MG content for wheat and rice flours, but had no significant effect for waxy rice flour.
Incorporation of MG into all flours significantly elevated the peak viscosity by about 0.9-2.6 folds when compared to non-MG
samples. Hot paste viscosity, breakdown and final viscosity for all flour mixtures significantly increased with increasing of MG
which ranged from 81-427, 37-559 and 152-463 RVU, respectively. Freeze-thaw stability measurement demonstrated that higher
level of MG in wheat and rice gel mixtures could decrease syneresis. However, MG had no effect on syneresis of waxy rice gel.
Presence of MG in flours alters the pasting properties and syneresis effect. It is suggested that higher viscosity and lower syneresis of
gels could be modified by MG.

Key words: Malva nut gum, pasting characteristic, freeze-thaw stability, wheat flour, rice flour, waxy rice flour.

1. Introduction relation to changes of quality such as paste viscosity

Rice, waxy rice and wheat are the main sources of
starches which differ significantly in composition,
morphology, thermal, rheological and retrogradation
properties. The physicochemical properties are
imparted by the starches to the aqueous systems and
their uniqueness in various food applications vary
with the biological origin. Starches or flours are an
important texture modifier in foods. They contribute
greatly to the textural properties of many foods and
have many industrial applications which act as a
thickener, colloidal stabilizer, gelling agent, bulking
agent, water retention agent and adhesive [1].
However, the problem of the applications of starch
or flour in food products may give different results in
Corresponding author: Yuthana Phimolsiripol, Ph.D.,
research fields: cereal and starch technology, product
development. E-mail: yuthana.p@cmu.ac.th.
and texture instability. This is due to native starches or
flours do not generally have ideal properties for the
preparation of food products [2, 3]. Unstable of gel
results in water separation after freezing and frozen
storages. Low temperature storage or repeated
freeze-thaw cycle during storage of starch-based
pastes gives rise to cryotropic gel formation, the final
products of which are in the form of sponge-like
textures. Repeating the freeze-thaw cycle enforces the
phase separation and ice growth, which leads to water
separation from the gel or syneresis [4]. Therefore,
hydrocolloids are usually added to starch-based foods
due to their desirable effect on the acceptability of
food products, because they provide several beneficial
effects with the improvement of food texture [5, 6] by
the retardation of starch retrogradation [7], increase of
moisture retention with the reduction of syneresis
 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour
resulting in extend the overall quality of the product
during the storage time [8, 9]. At the same time,
hydrocolloids have a variety of different structures,
including differences in branching, flexibility,
molecular weight ranges, and ionic charge, all of
which influence their behavior and the rheology of
their solutions. Some hydrocolloids such as guar gum
or locust bean gum are not pure polysaccharides and
contain protein, fiber, ash, etc. [10]. These results in
different modifications of rheological and textural
properties and improve the stability of food products
when hydrocolloids have been applied [11-13]. The
paste viscosity and gel texture of wheat flour
incorporated with fenugreek has greater hardness,
although the gels become softer with higher levels of
fenugreek content [14]. Furthermore, the effect of
hydrocolloids on the pasting characteristics is
dependent on types and concentrations of
hydrocolloids [15, 16]. Song et al. [17] reported that
gellan gum, xanthan gum, guar gum and gum arabic
increased apparent viscosity of rice starch dispersions.
Low concentration (0.05%) of hydrocolloids had no
effect on total setback viscosity appeared. Reduction
of retrogradation of tapioca starch is pronounced using
0.5% of xanthan gum [18]. In addition, syneresis
effect of rice starch can be decreased when starch
containing 0.5% of konjac glucomannan after
subjected to five freeze-thaw cycles [19].
Malva nut seed (Scaphium scaphigerum) contains a
large amount of mucilaginous substance and it has
been used as medicine for long time especially in
South East Asia. Malva nut gum (MG) is also
considered as hydrocolloids which linked with the
major carbohydrates including arabinose, galactose
and rhamnose. The molecular weight of MG is very
high, but is significantly reduced by the purification
processing (dialysis and decoloring). FT-IR
spectroscopy and methylation analysis show that MG
has a similar structure to gum arabic [20]. However,
MG is the impure polysaccharide which contained
6.72% protein, 3.83% crude fiber and 4.99% ash
299
content [21], and may affect the physicochemical
properties of starch-based mixtures. Purification
process by different concentrations of sodium
hydroxide is used to discard the protein content in the
MG. Furthermore, the alkaline extraction increases the
storage modulus of the gum by reduction of
galacturonic acid. Rheological properties were
modified when the gum was extracted with alkaline
solutions. Storage modulus increased with the
reduction of carboxylic bonds in the galacturonic
acids [22]. Due to high water absorption capacity of
the MG (ca 81 g per g dry sample) [21], the beneficial
aspect of MG is observed in the reduction of cooking
loss and improved textural properties of frankfurters
[23]. Therefore, the MG has a potential to use as a
thickening agent in protein-based or starch-based food
products.
Although some researchers have been reporting the
applications of MG in some starch-based food systems,
the information of MG effect on different flours is not
well known. Moreover, the interactions between
hydrocolloids and starches depend on the specific
starch or flour used [24]. The physiochemical
properties including pasting characteristics and
freeze-thaw stability of starch or flour can also be
altered by mixture interactions. Understanding the
interactions between flour-MG mix systems will be
useful for application of MG in the starch-based food
system in order to better physical and sensorial
properties. Therefore, the objective of this research
was to investigate the effect of MG replacement on
the pasting characteristics and freeze-thaw stability of
commercial wheat, rice or waxy rice flour systems.
2. Materials and Methods
2.1 Malva Nut Gum Preparation
Malva nut seeds were obtained from Chantaburi
province, Thailand. The MG was extracted following
the method of Phimolsiripol et al. [21]. The Malva nut
seeds were washed and soaked in water at the ratio of
1:80 (w/v) at 30 C for two hours to completely
300 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour
hydrate and swell the seed. The shell of the seed was
removed by hand and the jelly suspension was filtered
through a 60-mesh silk screen to remove the excessive
water. The mucilage was precipitated with three
volumes of 95% absolute ethanol and dried at 40 C
for 12 hours. Dried MG was milled using an ultra
centrifugal mill (ZM 200, Retsch, Germany) and then
sieved through a 80-mesh screen. The MG sample was
vacuum packed in laminated foil bags until further
analyzed.
2.2 Preparation of Mixtures
The mixtures of MG with wheat flour (United Flour
Mill Co. Ltd., Thailand), rice flour (Thai Wah Food
Products Public Co. Ltd., Thailand) or waxy rice flour
(Thai Wah Food Products Public Co. Ltd., Thailand)
were prepared by replacement technique. Wheat, rice,
waxy rice flours were replaced with MG at the
concentrations of 0, 0.5%, 1%, 2%, 3% and 5% (w/w)
of the flour weight. The ratios of dry powder mixture
of flour and MG corresponded to 100:0, 99.5:0.5, 99:1,
98:2, 97:3 and 95:5 w/w.
2.3 Pasting Characteristics of Flour-MG Mixtures
Pasting characteristics of the different mixtures were
examined using a Rapid Visco Analyzer (RVA-4D,
Newport Scientific, Warriewood, Australia). A
suspension of 10% wheat/MG powder in de-ionized
water (db, w/v) was heated from 50 C to 95 C at a
uniform rate of 12 C/min with constant stirring at 160
rpm. The sample was held at 95 C for 2.5 min, then
cooled to 50 C at a rate of 13 C/min, and held for 2
min. Total cycle time was 13 min using the standard
profile [25]. Pasting temperature (PT), peak viscosity
(PV), breakdown (BD), hot paste viscosity (HPV), final
viscosity (FV), and setback (SB) were recorded. The
viscosity values were reported in RVA units (RVU).
All measurements were performed in triplicate.
2.4 Syneresis Measurement by Freeze-thaw Cycles
Syneresis of each flour-MG gel mixture was
investigated by the modified method of Muadklay and
Charoenrein [18]. Samples of gelatinized gel mixtures
(obtained from the RVA determination above) of 0.5
mL were placed in 2 mL polycarbonate centrifuge
tubes with closed screw caps. The tubes were frozen
in a freezer chest at -20 C (MDF-U536d, Sanyo,
Japan) for 24 h and then thawed in the storage
chamber at 25 C for 1 h. The freeze-thaw cycle was
repeated for up to three cycles by 3-day intervals. Four
samples from each condition were centrifuged at
8,000 g in a refrigerated centrifuge for 15 min. The
supernatant was decanted and the residue was
weighed. The data were reported as the average of
four measurements. The percentage of syneresis was
quantified as the ratio of the weight of the supernatant
decanted to the total weight of the gel before
centrifugation and multiplied by 100.
2.5 Statistical Analysis
Analysis of variance (ANOVA) was performed to
find out if the effect of different concentration of MG
on the properties of flour mixtures were significant (P
< 0.05). Duncans multiple range test (DMRT) was
performed for multiple comparisons.
3. Results and Discussion
3.1 Pasting Characteristics
Incorporation of MG in wheat, rice or waxy rice
flours promoted a modification of the pasting
properties including pasting temperature and viscosity.
The pasting characteristics of wheat, rice or waxy rice
flour mixed with MG are presented in Fig. 1. The
pasting temperature values measured in the RVA for
wheat, rice or waxy rice flours mixed with 0, 0.5%,
1%, 2%, 3% and 5% of MG are demonstrated in Fig.
2. The PT values of the pastes significantly decreased
(P < 0.05) with increasing MG content for wheat and
rice flours, but had no significant effect (P 0.05) for
waxy rice flour. The PT values of mixtures were
ranged from 60 C-87 C. The gelatinization of starch
involved random entanglements of starch molecules
and the interaction of starch with water [26]. The
 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability 301
of Commercial Wheat, Rice or Waxy Rice Flour

(a) (b)

(c)
Fig. 1 Pasting characteristics of (a) wheat flour, (b) rice flour, and (c) waxy rice flour mixed with different concentrations of
MG.
302 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour

100 800
a
0% 0%
a a
a 0.5% 0.5%
a a
a 1% 1%
80 a 2% 2%
bb 3% 600 3% b b
Pasting temperature (oC)

b a a a a a
5% 5%

Peak viscosity (RVU)

b b
b
a c
60 c
a d
400 e
f
d
40
b e
c f
200 d
20 e e

0 0
W R GR W R GR
Fig. 2 Effect of different concentrations of MG on pasting Fig. 3 Effect of different concentrations of MG on peak
temperature of wheat (W), rice (R) or waxy rice (GR) flour. viscosity of wheat (W), rice (R) or waxy rice (GR) flour.
Mean values with different letters are significantly different Mean values with different letters are significantly different
(P < 0.05) for each flour. (P < 0.05) for each flour.

temperature at the onset of this rise in viscosity can be
considered as the starting point of gelatinization and is
defined as the PT in RVA test. Ragaee and Abdel-Aal
[27] found that the PT of wheat flour varied from
80 C-95 C depending on the types of flour and flour
mixture.
The most basic rheological characteristic of starch
paste is viscosity, which changes in broad range upon
applied shearing. The viscosities of pastes were
strongly affected by MG. Incorporation of MG in all
flours significantly elevated (P < 0.05) the peak
viscosity by about 0.9-2.6 folds when compared to
non-MG samples (Fig. 3). Rice flour had higher
viscosity than waxy rice and wheat flours. The PV of
wheat, rice and waxy rice incorporated with MG were
196-396, 204-729 and 346-665 RVU. This is
consistent with the results of Liu et al. [28] and Huang
et al. [29] who found that addition of hydrocolloid
altered the pasting behaviour. The increase in PV
could be accounted by a thickening gum that enhances
the forces being exerted on the starch granules in the
shear field and is attributed to the hydrocolloid
forming a network which entrapped granules. The
hydrocolloids could also form a strong entanglement
with the amylose released from starch granules [27,
29]. These phenomena were also found in the increase
in viscosity during pasting of waxy maize starch,
attributing the higher viscosities of the pastes
containing guar gum partly to the higher molecular
mass of the gum [30]. The mechanism of increasing in
PV for all flours incorporated with MG is due to high
water absorption of MG. It is responsible for a higher
synergistic effect with starch as similar to the results
of Brennan et al. [11] who found that the gums such
as guar gum or locust bean gum absorbed higher water,
resulting in greater viscosity of pastes. When
hydrocolloid is incorporated with starch paste, water
absorption of hydrocolloid plays an important role on
the viscosity. Hydrophilic hydrocolloid can bind with
water much faster than the starch granule. Thus, the
available water for starch granule swelling is reduced
and this leads to increase in the work required to move
swollen starch granules past each other in the more
viscous medium [31]. In addition, the alteration of
rheological properties of pastes depends on the
molecular mass and ranges of galactose and mannose
ratios. Higher branching degree of the gum could play
an important role in the interaction between the gum
and amylose molecules in starch [32]. The results
obtained in this study also indicated that the
non-starch components of flour may promote the
enhancement of MGs viscosity in synergism with
flours.
Swelling of granules accompanied by leaching of
 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability 303
of Commercial Wheat, Rice or Waxy Rice Flour

amylose increases the viscosity while granules may viscosity. The FV significantly increased (P < 0.05)
rupture during further shearing which results in with increasing in MG content for all flours (Fig. 6).
viscosity decrease. Hot paste viscosity, breakdown As the mixture cools, there is a decrease in kinetic
and final viscosity for all flour mixtures significantly energy, which allows the starch molecules to
increased (P < 0.05) with increasing of MG which re-associate and form network. The short-term
ranged from 81-427, 37-559 and 152-463 RVU, re-association results in textural changes of cooked
respectively as presented in Fig. 4. Leached amylose paste. Longer storage induces reversible
is more or less aligned in the direction of flow that re-crystallization of amylopectin, which increases the
contributes to the breakdown. The BD is correlated rigidity of the swollen granules embedded in the
with the stability of starch granule under the high continuous amylose network [32]. The SB values
shear conditions [25]. It is shown that increasing of were used to indicate the extent of short-term
MG ratio gave greater BD. Samples with 5% of MG retrogradation [4]. However, the reductions in the
showed the highest BD value for all flours (Fig. 5). setback from peak of all flours were promoted when
Final viscosity showed the similar trend with peak higher MG was added (Fig. 7). Greater effect on SB
600 800
a 0% 0%
0.5% 0.5%
1% a 1%
500 b
2% b 2%
3% 600 c 3%
Hot paste viscosity (RVU)

5%
Final viscosity (RVU)

c 5% d
400
e
a
d a a f
300 400 a
b b
e c
a d c b
e c e d
200 f
f
200 e e d
b
c
100 e e d

0 0
W R GR W R GR
Fig. 4 Effect of different concentrations of MG on hot Fig. 6 Effect of different concentrations of MG on final
paste viscosity of wheat (W), rice (R) or waxy rice (GR) viscosity of wheat (W), rice (R) or waxy rice (GR) flour.
flour. Mean values with different letters are significantly Mean values with different letters are significantly different
different (P < 0.05) for each flour. (P < 0.05) for each flour.
500 300
0% 0%
a
0.5% 0.5%
b ab
1% 200 1%
a
400 2% c 2%
3% 3%
Setback from peak (RVU)

5% 100 5%
Breakdown (RVU)

a a ab ab bc d
c
300 b
0
c
a a -100 e
200 d
e a a
f b
b
-200 c
b
c
100 d
d
ee c -300
d cdcd e

0 -400
W R GR W R GR
Fig. 5 Effect of different concentrations of MG on Fig. 7 Effect of different concentrations of MG on setback
breakdown of wheat (W), rice (R) or waxy rice (GR) flour. from peak of wheat (W), rice (R) or waxy rice (GR) flour.
Mean values with different letters are significantly different Mean values with different letters are significantly different
(P < 0.05) for each flour. (P < 0.05) for each flour.
304 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour

40
0%
35
0.5%
30 1%
% syneresis
25 2%
20 3%
5%
15
10
5
0
0 1 2 3 4
Number of freeze-thaw cycles
(a)
40
0%
35
0.5%
30 1%
% syneresis

25 2%
20 3%
5%
15
10
5
0
0 1 2 3 4
Number of freeze-thaw cycles
(b)
Fig. 8 Effect of different concentrations of MG on syneresis of (a) wheat gel and (b) rice gel.

was observed in waxy rice flour, but there only had
the small reduction in wheat flour.
Freeze-thaw cycles had a strong effect on the gel
stability as reflected by the increased syneresis when
the number of freeze-thaw cycles increased. The
syneresis from the gelatinized wheat or rice flour
pastes increased with increasing numbers of
freeze-thaw cycles (Fig. 8). Freeze-thaw stability test
showed that higher level of MG in wheat and rice gel
mixtures could decrease syneresis. After three
freeze-thaw cycles, incorporation of 5% MG reduced
syneresis by about 77% and 78% when compared to
non-MG samples for wheat gel and rice gel,
respectively. Higher water separation on thawing
indicates that the retrograded starch network or
structure was easily disrupted by ice crystal formation
[33]. However, MG had no effect on syneresis of
waxy rice gel. Overall, the results revealed that
retardation in syneresis was pronounced when higher
MG was used. Lower syneresis is probably due to
high intracellular and intermolecular hydrogen
bonding of hydrocolloid. The water-holding capacity
of hydrocolloid is the main functional property of
reduction in syneresis [34]. Ability of the gel to hold
moisture is dependent on the conformation and
composition of biopolymer systems [35].
4. Conclusions
The incorporation of MG into flours alters the
pasting characteristics and syneresis effect, depending
 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour
on types of flours. Higher paste viscosity and lower
gel setback were pronounced when MG was applied.
Pasting temperature decreased with increase MG
content. In addition, MG showed the effective agent
for the reduction of syneresis in wheat and rice flour
gels subjected to repeated freeze-thaw cycles. In this
study, application of 3%-5% of MG in wheat and rice
flour gels enhanced the freeze-thaw stability by about
45%-78% of syneresis reduction. However, it was not
detected in waxy rice gel. This result suggests that MG
can be applied as the useful additive for the quality of
starch-based food products in terms of viscosity
alteration and gel stability improvement after freezing
and thawing.
Acknowledgments
Author would like to thank the financial support
from the New Researcher Grant, Chiang Mai
University Research Fund.
References
[1] N. Singh, J. Singh, L. Kaur, N. Singh Sodhi, B. Singh
Gill, Morphological, thermal and rheological properties
of starches from different botanical sources, Food
Chemistry 81 (2003) 219-231.
[2] V.I. Lozinsky, L.G. Damshkaln, R. Brown, I.T. Norton,
Study of cryostructuration of polymer systems: XVI.
Freeze-thaw-induced effects in the low concentration
systems amylopectin-water, Journal of Applied Polymer
Science 75 (2000) 1740-1748.
[3] E. Armero, C. Collar, Antistaling additives, flour type
and sourdough process effects on functionality of wheat
doughs, Journal of Food Science 61 (1996) 299-303.
[4] R. Pongsawatmanit, S. Srijunthongsiri, Influence of
xanthan gum on rheological properties and freeze-thaw
stability of tapioca starch, Journal of Food Engineering
88 (2008) 137-143.
[5] R. Sharadanant, K. Khan, Effect of hydrophilic gums on
frozen dough: I. Dough quality, Cereal Chemistry 80
(2003) 764-772.
[6] T. Funami, S. Noda, M. Hiroe, I. Asai, S. Ikeda, K.
Nishinari, Functions of iota-carrageenan on the
gelatinization and retrogradation behaviors of corn starch
in the presence or absence of various salts, Food
Hydrocolloids 22 (2008) 1273-1282.
[7] T. Arunyanart, S. Charoenrein, Effect of sucrose on the
freeze-thaw stability of rice starch gels: Correlation with
305
microstructure and freezable water, Carbohydrate
Polymers 74 (2008) 514-518.
[8] P.D. Williams, L.N. Sadar, Y.M. Lo, Texture stability of
hydrogel complex containing curdlan gum over multiple
freeze-thaw cycles, Journal of Food Processing and
Preservation 33 (2009) 126-139.
[9] S. Simsek, Application of xanthan gum for reducing
syruping in refrigerated doughs, Food Hydrocolloids 23
(2009) 2354-2358.
[10] J.N. BeMiller, Pasting, paste, and gel properties of
starch-hydrocolloid combinations, Carbohydrate
Polymers 86 (2011) 386-423.
[11] C.S. Brennan, M. Suter, T. Luethi, L. Matia-Merino, J.
Qvortrup, The relationship between wheat flour and
starch pasting properties and starch hydrolysis: Effect of
non-starch polysaccharides in a starch gel system,
Starch/Starke 60 (2008) 23-33.
[12] J. Techawipharat, M. Suphantharika, J.N. BeMiller,
Effects of cellulose derivatives and carrageenans on the
pasting, paste, and gel properties of rice starches,
Carbohydrate Polymers 73 (2008) 417-426.
[13] Y. Zhou, D. Wang, L. Zhang, X. Du, X. Zhou, Effect of
polysaccharides on gelatinization and retrogradation of
wheat starch, Food Hydrocolloids 22 (2008) 505-512.
[14] C.S. Brennan, M. Suter, L. Matia-Merino, T. Luethi, G.
Ravindran, K. Goh, et al., Gel and pasting behaviour of
fenugreek-wheat starch and fenugreek-wheat flour
combinations, Starch/Starke 58 (2006) 527-535.
[15] P. Hongsprabhas, K. Israkarn, C. Rattanawattanaprakit,
Architectural changes of heated mungbean, rice and
cassava starch granules: Effects of hydrocolloids and
protein-containing envelope, Carbohydrate Polymers 67
(2007) 614-622.
[16] M. Huang, J.F. Kennedy, B. Li, X. Xu, B.J. Xie,
Characters of rice starch gel modified by gellan,
carrageenan, and glucomannan: A texture profile analysis
study, Carbohydrate Polymers 69 (2007) 411-418.
[17] J.Y. Song, J.Y. Kwon, J. Choi, Y.C. Kim, M. Shin,
Pasting properties of non-waxy rice starch-hydrocolloid
mixtures, Starch/Starke 58 (2006) 223-230.
[18] J. Muadklay, S. Charoenrein, Effects of hydrocolloids
and freezing rates on freeze-thaw stability of tapioca
starch gels, Food Hydrocolloids 22 (2008) 1268-1272.
[19] S. Charoenrein, O. Tatirat, K. Rengsutthi, M. Thongngam,
Effect of konjac glucomannan on syneresis, textural
properties and the microstructure of frozen rice starch
gels, Carbohydrate Polymers 83 (2011) 291-296.
[20] P. Somboonpanyakul, Q. Wang, W. Cui, S. Barbut, P.
Jantawat, Malva nut gum. (Part I): Extraction and
physicochemical characterization, Carbohydrate
Polymers 64 (2006) 247-253.
[21] Y. Phimolsiripol, U. Siripatrawan, C.J.K. Henry, Pasting
306 Effect of Malva Nut Gum on Pasting Characteristics and Freeze-thaw Stability
of Commercial Wheat, Rice or Waxy Rice Flour
behaviour, textural properties and freeze-thaw stability of
wheat flour-crude malva nut (Scaphium scaphigerum)
gum system, Journal of Food Engineering 105 (2011)
557-562.
[22] A. Srichamroen, V. Chavasit, Rheological properties of
extracted malva nut gum (Scaphium scaphigerum) in
different conditions of solvent, Food Hydrocolloids 25
(2011) 444-450.
[23] P. Somboonpanyakul, S. Barbut, P. Jantawat, N.
Chinprahast, Textural and sensory quality of poultry meat
batter containing malva nut gum, salt and phosphate,
LWT-Food Science and Technology 40 (2007) 498-505.
[24] V.M.F. Lai, A.L. Huang, C.Y. Lii, Rheological properties
and phase transition of red algal polysaccharide-starch
composites, Food Hydrocolloids 13 (1999) 409-418.
[25] K. Jangchud, Y. Phimolsiripol, V. Haruthaithanasan,
Physicochemical properties of sweet potato flour and
starch as affected by blanching and processing,
Starch/Starke 55 (2003) 258-264.
[26] L. Wang, Z. Yin, J. Wu, Z. Sun, B. Xie, A study on
freeze-thaw characteristics and microstructure of Chinese
water chestnut starch gels, Journal of Food Engineering
88 (2008) 186-192.
[27] S. Ragaee, E.S.M. Abdel-Aal, Pasting properties of starch
and protein in selected cereals and quality of their food
products, Food Chemistry 95 (2006) 9-18.
[28] H. Liu, N.A.M. Eskin, S.W. Cui, Interaction of wheat and
rice starches with yellow mustard mucilage, Food
Hydrocolloids 17 (2003) 863-869.
[29] C.C. Huang, P. Lai, I.H. Chen, Y.F. Liu, C.C.R. Wang,
Effects of mucilage on the thermal and pasting properties
of yam, taro, and sweet potato starches, LWT-Food
Science and Technology 43 (2010) 849-855.
[30] S. Satrapai, M. Suphantharika, Influence of spent
brewer's yeast -glucan on gelatinization and
retrogradation of rice starch, Carbohydrate Polymers 67
(2007) 500-510.
[31] M. Alloncle, J. Lefebvre, G. Llamas, J.L. Doublier,
Rheological characterisation of cereal
starch-galactomannan mixtures, Cereal Chemistry 66
(1989) 90-93.
[32] M.J. Miles, V.J. Morris, P.D. Orford, S.G. Ring, The
roles of amylose and amylopectin in the gelation and
retrogradation of starch, Carbohydrate Research 135
(1985) 271-281.
[33] M.H. Lee, M.H. Baek, D.S. Cha, H.J. Park, S.T. Lim,
Freeze-thaw stabilization of sweet potato starch gel by
polysaccharide gums, Food Hydrocolloids 16 (2002)
345-352.
[34] C. Ferrero, M.N. Martino, N.E. Zaritzky, Corn
starch-xanthan gum interaction and its effect on the
stability during storage of frozen gelatinized suspension,
Starch/Starke 46 (1994) 300-308.
[35] Y.M. Lo, K.L. Robbins, S. Argin-Soysal, L.N. Sadar,
Viscoelastic effects on the diffusion properties of curdlan
gels, Journal of Food Science 68 (2003) 2057-2063.
Journal of Agricultural Science and Technology B 2 (2012) 307-311
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Use of High Content Fish Silage Wet Food in Final

Growth out of Pac (Piaractus mesopotamicus,
Holmberg 1887) in Northeast Argentina

Gustavo Wicki, Oscar Galli Merino, Pablo Cal and Facundo Sal
Centro Nacional de Desarrollo Acucola (CENADAC), Direccin de Acuicultura-SAGPyA, Paseo Coln 982, Anexo Pesca 1063,
Bs.As., Argentina

Received: September 9, 2011 / Published: March 20, 2012.

Abstract: The effect of two diets containing fish acid silage were studied, the first one offered as dry diet and the second one as a
wet diet, in final culture cycle of Pac. Previous experiences showed that diets without fishmeal maintain same growth rate in semi
intensive system, compared to diets containing 8% and 20% fish meal. Feeding trials were carried out in CENADAC (2732 S,
5830 W) in a semi intensive pond culture system for 122 days. Two experimental feed including fish silage, a liquid product
obtained of fish residues treated by enzymes action, were formulated. Fish silage was prepared by mincing and blending fish viscera
with sufficient formic acid (2.11% v/w) to attain a stable pH of 3.5. Fish were daily fed initially with the dry ration at 1.3% bw/d and
at a rate of 0.55 g of protein/100 g of fish with the wet ration. At the end of the experience the average weight observed for dry fed
fish was 1,513.5 g and 1,560.3 g for those wet fed. The daily growth was 3.69 g/d for the first group and 3.9 g/d for the second one.
The final FCR was 2.72 and 3.01 for each group without significant differences. These results showed that it is possible to grow out
pac with wet self made feed without fish meal and with similar growth that those fed dry diet.

Key words: Piaractus mesopotamicus, feed, fish acid silage.

1. Introduction exchange or supplementary aeration is approximately

16 months [2, 3].
Pac (Piaractus mesopotamicus) is a fast growing
The pac is considered an omnivorous fish with
fish species inhabiting the La Plata basin in South
trend to herbivorous [4], for that reason it is feasible to
America. However the species disappeared over 20
suppose that it can be fed with a low content or even
years ago from the Uruguay River and more recently
without fish meal, which would represent a lower
from a great portion of the Paran river. The
productive cost.
disappearance has been mainly due to environmental
Long term pond feeding trials carried out at the
changes in the basin related to dam construction [1].
CENADAC found out that diets with 8% fish meal or
Since year 2000, pac production has increased
12% fish acid silage and without fish meal had similar
from 70 to 600 TM/year in Northern Argentina. Large
growth that those with a inclusion of 20% fish meal [5].
size pac of 1.2 kg/pc is preferred by local consumers.
The use of wet food is a good choice for small fish
Reported growth period required to reach this market
farmers (lower than 10 TM fish/year) because it can
size under semi-intensive pond culture conditions at
be made daily according to requirement.
low stocking densities (0.2 ind/m2) and no water
The fish acid silage can substitute totally or
partially fish meal, a high cost ingredient, and
Corresponding author: Gustavo Wicki, M.Sc., research
fields: aquaculture, nutrition, native species culture techniques promotes the utilization of fish viscera and other fillet
development. E-mail: guswicki@gmail.com.
308 Use of High Content Fish Silage Wet Food in Final Growth out of Pac
(Piaractus mesopotamicus, Holmberg 1887) in Northeast Argentina

residues. Animal residues are a great nutrient and Table 1 Feed rations tested during the feeding trial
(values expressed as percentage).
energy reservoir that can be transformed in a product
Ingredients Dry food (A) Wet food (B)
of high biological value such as in the fish acid silage.
Meat meal 18 10
Simple equipment is required to make fish acid Fish silage 20 40
silage, a meat grinder and a mixer. Same equipment is Soy meal 41 30
used to make the wet diet. Rice bran 17 16
(PO4)2Ca3 2 2
This type of diets has been widely tested within
Vitamin 1 1
different species, different levels of acid silage Salt 1 1
(between 40%-80%) in formula and different types of Total 100 100
organic and inorganic acids [6-8]. Crude protein 32 25
Microbiological quality of fish acid silage has been Moisture 10 33

proved by several authors [9, 10].
The objective of this study was to evaluate
performance of a wet diet compared to a dry one, both
with inclusion of fish acid silage.
2. Materials and Methods
Feeding trials were carried out in CENADAC
(2732 S, 5830 W) in a semi intensive pond culture
system for 122 days, with three ponds (300 m2 each
one) allocated per dietary treatments.
Two experimental feed including fish silage were
formulated, the first one (treatment A) offered as dry
pellets contained, fish silage (20%), meat meal (18%)
soybean meal (41%), and rice bran (17%). The second
one (treatment B) offered as wet pellets contained,
fish silage (40%), meat meal (10%), soybean meal
(30%), and rice bran (16%). The composition of both
diets is showed in Table 1.
Fish silage was prepared by mincing and blending
fish viscera with sufficient formic acid (2.11% v/w) to
attain a stable pH of 3.5 according to Manca and
Carrizo [11]. Chemical composition of the silage was
as follows: moisture 76.8%, crude protein 12.8%, fat
8.6%, and ash 1.6%. Stabilized acid silage was stored
at room temperature inside sealed plastic flasks until
used.
Water quality variables were monitored twice daily,
including temperature, dissolved oxygen and pH
during the early morning and late in the afternoon
prior to feeding.
Fish were fed once daily, six days per week,
initially at 1.3% of their body weight per day for those
fed with dry ration and at a rate of 0.55 g of
protein/100 g of fish per day for those fed with wet
ration.
The way to estimate the amount of wet feed is
shown in Box 1.
Food conversion ratio (FCR) was calculated in dry
basis in order to compare both diets.
Box 1 Step by step wet feed calculate mode:
Q = protein content (g)/100 g of fish, so 0.55 g
prot/100 g fish = 5.5 g prot/kg fish;
B = biomass of fish (kg), e.g.: 150 kg of fish;
A = amount of feeding protein = Q B (g) e.g.: 5.5
g/kg 150 kg = 825 g of feeding protein;
PF = protein in feed (g/kg feed) e.g.: 25% = 250 g
of protein/kg of feed;
DD = daily diet = A/PF = Kg feed;
e.g.: 825 g/250 g/kg = 3.3 kg feed.
Ten percent of fish population was sampled on a
monthly basis, to determine fish growth, fish health
and to recalculate required feeding rate.
All the fish were weighted at the end of the
experiment. Survival rates and overall yield were
determined, as well as feed conversion ratio (FCR =
food offered/weight gained) and daily growth (DG =
(final weight-initial weight)/time).
Statistical data analysis was performed and
treatments groups were compared using one-way
ANOVA, according to Hintze [12].
 Use of High Content Fish Silage Wet Food in Final Growth out of Pac 309
(Piaractus mesopotamicus, Holmberg 1887) in Northeast Argentina

3. Results and Discussion season [5, 13].

3.1 Water Quality Results 3.2 Growth Response

Mean water temperature during the feeding trial
was 26.5 C (Fig. 1), with maximum value found on
January of 34.6 C, and a minimum at the end of
summer, March of 21.7 C. Mean dissolved oxygen
resulted 4.5 mg/L. Dissolved oxygen levels reached
maximum and minimum peaks during January,
ranging from 10.2 mg/L in the afternoon to 1.66
mg/L in the early morning. Water pH values
displayed an average of 7.71 over the culture cycle,
with slight daily and seasonal variation ranged from
7.50 to 7.88.
Results show no different from those obtained at
the same site in different years of typical summer
40
35
30
DO(mg/L) T(C)
Average final fish weights were 1,513 g for those
fishes that consumed dry feed and 1,560 for the
group fed with wet diet; differences between
treatments were statistically significant (P = 0.001).
Fish growth during the feeding trial with both diets is
shown in Fig. 2.
Average daily growth resulted of 3.69 g/day for the
dry feed group and 3.90 g/d for wet diet group,
without statistically significant differences (P = 0.5).
Previously reported studies ranged between 2.62 g/d
[14], and 5.65 g/d [5] during final growth phase.
Present values are located between those reported by
these authors.
Tprom

Tmax
25 Tmin
20 DOprom
15 DOmax
10 DOmin
5
0
Nov08 Dec08 Jan09 Feb09 Mar09
Fig. 1 Mean dissolved oxygen and water temperature registered during culture cycle.

1700

1600

1500
Weight(g)

1400
TreatmentA
1300
TreatmentB
1200

1100

1000
november december january february march
Fig. 2 Growth of pac with different diets.
310 Use of High Content Fish Silage Wet Food in Final Growth out of Pac
(Piaractus mesopotamicus, Holmberg 1887) in Northeast Argentina

Galli Merino et al. [13] reported a daily growth of Table 2 Growth performance of Pac fed different diet,
dry (A) and wet (B).
3.7 g/d using commercial feed composed only with
Treatment A Treatment B
vegetables ingredients, similar results were obtained
Initial weight (g) 1,062.84 1,084.32
with the dry feed in this experiment. Final weight (g) 1,513.50 1,560.34
The high percentage of soybean meal (41%) used Survival (%) 100 100
in treatment A and the antinutritional factors Culture period (days) 122 122
Days fed (days) 78 78
contained in this ingredient may explain the lower
final harvest (kg/ha) 2,572.95 2,575.16
growth (although not statistically significant). FCR 2.72 3.01
Perez et al. [15] in an in vitro digestibility assay DG (g/d) 3.69 3.90
with pac, found a high sensitivity to inhibitors
Good quality pellets are obtained with 40% fish
present in soybean meal. Alkaline proteases were
acid silage included in feed.
highly inhibited in presence of moderated
Feeding rate must be adjusted in future studies in
concentrations of soybean extract.
order to obtain lower FCR.
The survival rate was 100% and the harvest at the
end of the experience resulted of 2.57 TM/ha for both References
treatments (Table 2). Wicki et al. [3] reported a final
[1] R. Quiros, The La Plata river basin: International basin
production of 2.89 TM/ha for a two summer pac development and riverine fisheries, in: Second
culture at 0.2 ind/m2 density. The difference was due to International Symposium on the Management of Large
the lower density used in the present experiment (0.17 River for Fisheries (LARS2), Phnom Penh, Kingdom of
Cambodia, 2003, p. 43.
ind/m2)
[2] G. Wicki, Cultivo y produccin de Pac (Piaractus
Average food conversion rate (FCR) resulted of 2.72 mesopotamicus): Incidencia de dos dietas de diferente
for treatment A and 3.01 for treatment B, without composicin y de la densidad de siembra en cultivo
statistically significant differences (P = 0.39). The FCR semi-intensivo, Tesis de Maestria en acuicultura, Escuela
para graduados, Fac. de Agronoma, Univ. de Buenos
obtained resulted higher than those reported by Wicki
Aires, Argentina, 2003, p. 81.
et al. [14] of 2.37 and Wicki and Luchini [5] of 1.96, [3] G. Wicki, L. Luchini, L. Romano, S. Panne, Stock
both using dry feed and a lower feeding rate (1%). densities, growth and survival for Pac (Piaractus
Value of 2.66 was reported by Galli Merino et al. [13] mesopotamicus), World Aquaculture Magazine 40 (2009)
51-54.
with commercial feed and the same system culture.
[4] R.A. Ringuelet, R.H. Aramburu, A. Alonso de Aramburu,
Average FCR obtained with wet diet resulted higher Los peces Argentinos de agua dulce, CIC, Buenos Aires,
compared with those reported by Toledo et al. [7] of Argentina, 1967, p. 248.
1.78. This author suggests a feeding rate of 0.43 g [5] G. Wicki, L. Luchini, Development of practical diets for
Pac, a south American fish species, International
CP/100 g fish for Clarias gariepinus. High feeding Aquafeed 7 (2004) 23-29.
rates could be the response of high FCR with little [6] J. Llanes Iglesias, J. Toledo Perez, J. Lazo de la Vega,
weight gain. Mximo porcentaje de ensilado qumico de pescado en la
dieta de Clarias gariepinus (Burchell, 1822), Redvet,
4. Conclusion Revista electrnica de veterinaria, VIII, No. 9, 2007.
[7] J. Toledo, A. Botello, J. Llanes, Evaluacin de los
Pac accepts wet diet easily, with weight gain ensilados qumicos de pescado en la alimentacin de
compared to dry diet. Clarias gariepinus (Burchell, 1822), Redvet, Revista
electrnica de veterinaria, VIII, No. 9, 2007.
In future experiences it will be determined if the
[8] M.A. Gomes Pimenta, O.M. Moraes, P.V. Logato, C.
better growth obtained with wet diet has been due to Pimenta, T.A. Freato, Desempenho produtivo e
better feed quality of wet diet or higher feeding rates. digestibilidade pela Tilapia do Nilo alimentada com
 Use of High Content Fish Silage Wet Food in Final Growth out of Pac
(Piaractus mesopotamicus, Holmberg 1887) in Northeast Argentina
dietas suplementadas com niveis crescentes de silagem
acida de pescado, Cienc. Agrotec. 32 (2008) 1953-1959.
[9] W.R. Boscolo, A.M. Dos Santos, M. Buzanello, C.
Fieden, F. Bittencourt, A.A. Signor, Avaliacao
microbiolgica e bromatolgica da silagem acida obtida
de residuos da industria de filetagem de tilapia do nilo
(Oreochromis niloticus), Ciencias Agrarias Londrina 31
(2010) 515-522.
[10] M.T. Garcia de la Cruz, M. Vinjoy Campa, G. Cabrera
Morales, Calidad microbiolgica de diferentes dietas
confeccionadas a base de ensilado de pescado, Acuacuba
3 (2001) 34-38.
[11] E. Manca, J.C. Carrizo, Informe final de produccin y
utilizacin de ensilados en la formulacin de dietas, Proy
DNA/INIDEP, expte 4961, Argentina, 2002, p. 8.
[12] J.L. Hintze, Number Cruncher Statistical System (NCSS,
2000),Version 6.0. graphics, Kaysville, Utah, USA, 1998.
[13] O. Galli Merino, G. Wicki, F. Sal, Utilizacin de un
311
alimento basado en ingredientes de origen vegetal en la
ltima fase de engorde de pac (Piaractus
mesopotamicus), in: II Conferencia Latinoamericana
Sobre Cultivo de Peces Nativos, Bs, As, Argentina, 2009,
p. 96.
[14] G. Wicki, F. Rossi, S. Martin, S. Panne, L. Luchini,
Engorde final de pac (Piaractus mesopotamicus) con
raciones basadas en subproductos de maz, girasol, y
ensilado cido, in: Desarrollo y utilizacin de ensilado
cido como componente de alimento para peces.
FAO-SAGPyA, Buenos Aires, Argentina, 2007, pp. 45-58.
[15] J.J. Perez, G. Wicki, F. Moyano, F. Alarcon, Evaluacin
del efecto de inhibidores de proteasa presentes en
ingredientes vegetales utilizables en piensos para dos
especies pisccolas cultivadas en Argentina, pac
(Piaractus mesopotamicus) y pejerrey (Odontesthes
bonariensis), in: II Congreso Iberoamericano Virtual de
Acuicultura (CIVA), 2003, pp. 442-454.
Journal of Agricultural Science and Technology B 2 (2012) 312-328
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Long-term Decomposition Process of the Leaf Litter,

Carbon and Nitrogen Dynamics under Different Forest
Management in the Sierra de Francia, Salamanca, Spain
Sergio Salazar1, Luz-Estela Snchez2, Purificacin Galindo3 and Ignacio Santa-Regina1
1. Institute of Natural Resources and Agrobiology of Salamanca, IRNASA-CSIC, Cordel de Merinas 40-52, Salamanca 37008, Spain
2. Department of Mathemaics, University of Antioquia, Colombia
3. Department of Statistics, University of Salamanca, Spain

Received: August 31, 2011 / Published: March 20, 2012.

Abstract: The dynamics of organic matter, carbon and nitrogen were studied during leaf decomposition over a three annual cycles
period using the litterbag method at three permanently untilled and unfertilised plots. Our hypothesis is to determine if each litter
type influences the decomposition and C and N releases from each other litter type (three plots and two species). The main objective
of this study is the comparison of decomposition dynamics in a climax forest with respect to that occurring in the chestnut managed
plots on terrain suitable for oak to have further insight into the recycling of above ground organic matter and these two associated
bioelements. As for the loss of dry matter, two-way analysis of variance, involving the treatment and year factors, revealed the
existence of significant differences only for year, while no differences were observed for either treatment or interaction. At the end of
the 2.4 years of the study, most of the treatment assays higher nitrogen concentrations than the initial ones were recorded. These three
treatments the samplings prior to the last one revealed nitrogen concentrations higher than the initial ones, such that it may be
concluded that the trend was towards an increase in concentration.

Key words: Litter decomposition, mass loss, carbon and nitrogen release, litter bags, forest ecosystems, Castanea sativa, Quercus
pyrenaica.
.

1. Introduction decomposition since they are controlled by the

interaction among two types of component: on one
The decomposition of organic matter is the result of
hand, abiotic factors such as the climate and
several processes acting in conjunction: the
mineralogy of the substrate [2], on the other hand, by
fragmentation of plant material carried out by biotic
biotic factors, such as the chemical composition of the
(activity of the fauna) and abiotic (impact of rain
leaf litter [3-8], and the composition and activity of
drops, etc.) processes; the leaching of water-soluble
soil organisms [9, 10]. These elements do not all have
components such as sugars, organic acids, amino acids
the same importance, since they act at different
and other simple molecules [1], and the respiratory
spatio-temporal scales. In general, it has been
processes of decomposing organisms (mainly bacteria
established that they act in the following order of
and fungi), which reduce and mineralize the organic
importance: climate > chemical composition > soil
matter to basic inorganic molecules, such as ammonia,
organisms [11], because the climate is the most
phosphate, CO2 and water. It is difficult to determine
important regulator of the metabolism of bacteria and
the relative contribution for each of these processes to
fungi. Considering large geographic scales, with a
Corresponding author: Ignacio Santa-Regina, researcher, broad degree of climatic variability, the climate is the
research fields: plant ecology, soil biogeochemistry and soil
biology. E-mail: ignacio.santaregina@irnasa.csic.es.
dominant regulatory factor (represented above all by
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics 313
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

mean annual temperature and evapotranspiration), terrain suitable for oak to have further insight into the
while the factors referring to leaf litter quality have recycling of above ground organic matter and these two
much lower predictive values [11-14]. Many authors associated bioelements. This study was carried out over
have detected correlations with different variables 2.4 year-period of experimentation, subject to the
(grown in relation to its initial litter quality) [15], and semiarid climate of the west-central Spain. This was
most have found that the variability explained by the done by incubating different leaf litter samples of
combination of two or three factors is greater than that chestnut and oak at the three plots.
accounted for by each of them individually (sun and
2. Material and Methods
shade leaves as affected by chemical composition) [6].
Also, combined temperature and precipitation [16] 2.1 Field Site and Experimental Set-up
and/or relative humidity [17]; Sariyildiz et al. [18] The research was carried out in the Sierra de Francia
concluded that mean annual temperature and mean (Honfra area), Salamanca province (Spain). Three
annual rainfall account for between 72% and 87% of permanent no-till and unfertilised plots were chosen:
the variance. the first in a chestnut paraclimax coppice (CC),
Sweet chestnut (Castanea sativa Mill.) stands are (403521 N, 55630 W), the second in a chestnut
very common around the western Mediterranean Basin. orchard (CO) and the third in a oak climax forest,
Formerly managed as coppices, these stands were adjacent to the CO plot (403437 N, 5571 W). The
regularly clear-cut every 15-25 years according to their densities of trees are (tree ha-1): 1,892, 382 and 2,960
productivities under several local conditions. However, for CC, CO and Oak, respectively, as shown in Table 1.
chestnut coppice management has now been more or The study area is dominated by granite substrates,
less abandoned, although chestnut coppices still cover and the soil pH is generally acid, although with
fairly large areas in mountainous Mediterranean areas. intercalations of limestone, there are enclaves of
Our hypothesis is to determine if each litter type vegetation typical of the Euro-Siberian region, with
influences the decomposition and C and N releases taxa such as: Ilex aquifolium L., Aconitum napellus
from each other litter type (three plots and two species). ssp., Castellanum L., Actaea spicata L., Monotropa
The main objective of this study is the comparison of hypopitys L., Atropa belladona L., Hypericum
decomposition dynamics in a climax forest with respect montanum L., Neottia nidus-avis (L.) Rich., Paris
to that occurring in the chestnut managed plots on quadriflia L., Corylus avellana L., etc.
Table 1 General characteristics of stands selected in Honfra forest area.
Stand C. sativa coppice (CC) C. sativa orchard (CO) Q. pyrenaica (Oak)
Altitude (m.a.s.l.) 1015 950 950
Soil type Umbric regosol Umbric leptosol Umbric leptosol
Leaf area index (m2 m-2) 2.9 3.1 2.5
Long term mean P (mm) (annual rainfall) 1,590 1,530 1,530
Mean annual temperature (C) 10.8 11.1 11.1
Range 5.35-19.3 18.9-22.7 5.2-23.6
Diameter at breast height (cm) 12.90 1.7 b 20.40 3.0 c 11.60 1.5 b
Tree age (years) 70 85 75
Tree height (m) 15.3 1.3 c 8.90 0.8 a 12.2 1.0 b
Shoot density (shoot ha-1) 1,892 100 b 382 30 a 2,960 125 c
Basal area (m2 ha-1) 28.40 8 c 18.50 4 a 26.50 7 c
Slope (%) 20 18 18
Confidence intervals P = 0.05. For each site, mean values in the same column followed by different letters are significantly different.
314 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
Mean annual temperature ranges around 10.0 C
[19], here, any differences were hardly observed
among the values obtained over the 2.4 years of study
(2001-2004) or owing to their closeness among the
three plots studied. Maximum monthly temperatures
were recorded in July and August, the values ranging
between 20 C and 22 C, and minimum temperatures
were recorded in January, a minimum value of -4 C
being observed at the oak stand in January, 2003.
In order to collect the material from the litter fall,
three groups of ten 0.24 m2 boxes were randomly
distributed in each plot, and the litter was collected
once a month and its components separated into
different plant organs.
Top-soil (0-10 cm) were sampled, ten, five and five
soil samples were collected respectively from each
soil profile in coppice, oak and chestnut orchard plots
respectively, and were taken three samples at each
sampling point, at the same time, to know the soil
characteristics.
Leaf decomposition dynamics were assessed in
litterbags of 400 cm2 surface area, with a mesh
diameter of 1 mm. Each litterbag contained 10 g of
freshly fallen leaves of the dominant tree species. In
total, 45 litterbags, 15 from each species at each of the
3 sites, were placed within the soil litter layer, from
the season of peak litter fall, distributed into three
groups (high, central and low lying part of each stand),
corresponding to the season of peak litter fall. Over
2.4 years, three litterbags from each location (one
from each species) were collected every two months,
and their contents were cleaned, weighed, ground, and
analysed. Additionally, from each site, litter samples
were collected from a 50 50 cm area of the ground
to determine the indices of natural decomposition in
the three plots.
2.2 Laboratory Methods
The samples were taken to the laboratory, where
they were dried in a Memmert desiccation oven at
80 C to determine their moisture content and were
then ground for later analysis.
Organic carbon was determined on oven-dried
samples using a Wsthoff Carmograph 12 device.
This device combusts 0.03 g of a ground, dried sample
in an oxygen atmosphere at 950 C-1,000 C. The
stream of gases produced passes to a secondary
combustion oven of copper oxide, where the CO
possibly formed is oxidised to CO2. The stream of
gases is then aspirated by different pumps through
different absorbents of SO2 and H2O and introduced in
the measuring cell, where the carbon dioxide, which is
formed equivalently by the combustion of carbon, is
mixed with NaOH for its reaction. The variation in the
electrical conductivity of the reagent is the parameter
that indicates the carbon content.
Determination of the total nitrogen content was
performed using a 3 Bran Luebbe Autoanalyzer.
Mineralization of organic nitrogen was accomplished
with wet digestion, for which 0.1 g of a ground
homogenised sample was weighed using a Sartorius
precision balance, transferred to a crucible, and
concentrated sulphuric acid plus a small amount of
SO4K2/Se/SO4Cu catalyst were added. Once digestion
had been completed, the volume was brought up to
100 mL, and 5 mL aliquots introduced into the A3
autoanalyzer, where the sample reacted with salicylate
and sodium hypochlorite to produce a blue compound
which was measured at 600 nm. Sodium nitroprussiate
was used as the catalyst.
For the estimation of litter dynamics, we used the
coefficient K proposed by Salazar [20], which relates
the humus and the aboveground litter (A and F),
where A is the leaf litter returning yearly to the soil,
and F is the leaf litter accumulated on the forest soil
before the period of the litter fall (to measure this, all
of the material of the holorganic horizon contained in
a 0.5 0.5 m frame was collected). Fifteen samples
per plot were taken [20]. K is a constant for a given
ecosystem and is defined as Eq. (1):
K = A/(A + F). (1)
The weight losses of the annual leaf litter
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
production (P) can be established as Eq. (2):
P = AK (2)
Calculation of the decomposition coefficient K0 (21)
is defined by the relation as Eq. (3):
K0 = A/F (3)
The parameter Kd, coefficient of accumulation of
leaf litter or total litter, was also determined as Eq. (4):
Kd = (A P)/A (4)
Data were subjected to a one-factor statistical
analysis of various algorithms (ANOVA). The
regression curves were also established according to
the best r.
2.3 Fitting of Regression Curves
Linear regressions were performed with the natural
logarithm of the mean dry mass remaining at each time
to calculate Kt, a constant of the overall fractional loss
rate for the study period, following the Eq. (5):
Ln (Xt/X0) = Kt. (5)
where Xt and X0 are the masses remaining at time t
and time zero respectively [21]. Both masses
remaining on the soil were calculated immediately
before the annual litterfall peak.
The loss of remaining mass versus the time elapsed
during the decomposition process of the leaves
contained in the bags was fitted with the following
simple exponential model as Eq. (6):
(Xt = X0 e-kt) (6)
In all cases, a significant degree of fitting was
obtained. The r2 values of the simple exponential
equation ranged between 0.95 (coppice chestnut
leaves at the oak stand) and 0.87 (oak leaves at the
chestnut coppice). This indicates that the negative
simple exponential model accounts for 95% and 87%
of the variance in the loss of mass with time.
2.4 Statistical Methods
To perform the different statistical tests, the SPSS
14.0 program was employed. The statistical
differences in the mean values of the parameters
studied in this work were estimated with one and
two-way ANOVA. When significant differences were
detected with these analyses of variance, post-hoc
315
tests with the LSD test (P < 0.05) were carried out to
determine where those significant differences lay.
Prior to the analysis of variance, the normality and
homoscedasticity of the data were checked with the
Shapiro-Wilks test.
To determine the relationship between weight loss
and the climatic variables, Pearson correlation
analyses were performed (P < 0.05). Linear regression
was also performed to determine the relationship
between the remaining mass and the concentrations of
carbon and nitrogen.
3. Results
3.1 Dry Matter Loss and Decomposition Rates
Approximately 60% of the dry matter of chestnut
and oak leaves confined in the litter bags at the three
plots was lost over the 890 days of the study (Fig. 1).
In order to check whether there were significant
differences in the loss of dry mass during the process
of decomposition of the three litter types (treatments)
used across three different sites, a two-way analysis of
variance was carried out, and the result was: treatment
and year coppice chestnut leaves at the chestnut
orchard (P < 0.001, significantly differences at P <
0.05, according to LSD test).
The parametric two-way analysis of variance only
detected significant differences for the year factor.
This confirms that the remaining masses of
decomposing leaves were significantly greater in the
first year compared to the second one in the nine
treatments (P = 0.000, significantly differences at P <
0.05, according to LSD test), but no differences in
mass loss were observed between sites or litter types.
The decomposition rate (k) of each type of leaf at
the different sites was calculated with the simple
exponential model [22] from the remaining masses.
The results are shown in Table 2. The decomposition
rate (K0) each litter type at the different sites, over
the full period, ranged between 0.33 (in the treatment
with oak leaves at the chestnut coppice) and 0.50
(chestnut leaves in the oak stand) (Table 2). As occurred
316 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

120,0
CO: y = 84,422e-0,449x R2: 0,94 CO
100,0 CC: y = 83,743e-0,299x R2 : 0,88 CC
Oak: y = 86,572e-0,435x R2 : 0,94
R em a inder w eig ht %

Oak
80,0

60,0

40,0

20,0

0,0
0 121 245 360 457 586 698 824 Days
winter spring summer winter spring summer winter spring
(a)

120,0 CO: y = 81,406e-0,4243x R2 : 0,94 CO

CC: y = 81,574e-0,3023x R2 : 0,87
100,0 CC
Oak: y = 85,861e-0,4588x R2 : 0,94
Remainder weight %

Oak
80,0

60,0
Oak
40,0

20,0

0,0
Days
0 121 245 360 457 586 698 824
winter spring summer winter spring summer winter spring
(b)
120,0 CO
CO: y = 83,855e-0,395x R2 : 0,95
CC
100,0 CC: y = 82,714e-0,328x R2 : 0,92
Oak
Oak: y = 85,686e-0,388x R2 : 0,91
R em a inder w eig ht %

80,0

60,0

40,0

20,0

0,0
0 121 245 360 457 586 698 824 Days
winter spring summer winter spring summer winter spring
(c)
Fig. 1 Dry mass loss during decomposition of leaves of orchard chestnut (CO), coppice chestnut and oak in the (a) chestnut
orchard, (b) the chestnut coppice, and (c) the oak forest (n = 3).
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics 317
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

Table 2 Decomposition rates (K0, year-1) of the different leaves calculated from the mass remaining according to the simple
exponential model.
C. sativa orchard (CO) C. sativa coppice (CC) Q. pyrenaica (Oak)
Year
CO CC Oak CO CC Oak CO CC Oak
2001 0.82 0.86 0.77 0.98 0.94 0.89 0.93 0.96 0.87
2002 0.58 0.63 0.46 0.63 0.60 0.48 0.51 0.54 0.48
2003 0.44 0.46 0.37 0.48 0.46 0.33 0.50 0.43 0.38

with the loss of dry mass, the parametric two-way
analysis of variance (treatment and year) only detected
significant differences for the year factor (F = 12.48, P
= 0.000, df = 2), but no differences were found for
either treatment or for the interaction treatment and
year (F= 163.1, P = 0.000, df = 2). Significantly
differences at P < 0.05, according to LSD test),
assuming that the decomposition rates collected in the
last sampling made do not change until the end of the
decomposition process, the theoretical times required
for 50%, 95% and 99% of the leaf litter to decompose
were calculated (Table 3).
3.2 Decomposition Indices
Several decomposition indices were calculated,
considering the total litter (leaves, branches and
reproductive structures) at each plot or only the leaves
litter (Table 4). In all three cases, considering only the
leaves were the decomposition indices higher than
those considering the total leaf litter. Regarding total
litter, the highest K and Ko decomposition coefficients
were estimated for the chestnut coppice, followed by
the oak stand and finally the chestnut orchard.
However, when only the leaves were considered, the
K and Ko decomposition coefficients at the two
chestnut stands were equal, and higher than those
found for the oak stand. To determine the influence of
the climatic variables on leaf litter decomposition,
correlations between the decomposition rates and
precipitation, mean temperature, maximum
temperature, minimum temperature,
precipitation/mean temperature ratio and the moisture
content of the leaf litter were determined, and found to
be significant for moisture content of the litter related
to all the litter leaves, and the rainfall related to oak
litter leaves placed in the chestnut coppice stand,
according to the coefficient of Spearman (Table 5).
3.3 Dynamic of Carbon and Nitrogen in Decomposing
Leaves
In the different treatments carried out, a decrease in
carbon concentrations was observed throughout the
almost two and a half years of the treatment. This
decrease was slightly greater in the first sampling.
Only in the treatments carried out with oak leaves at
the chestnut orchard and the oak leaves at the chestnut
coppice and the oak leaves at their own plot did we
fail to observe this decrease in the first sampling. In
the rest of the process, the trend was towards a
decrease in the concentration of carbon, although in
the final stages some increases were detected (Fig. 2).
The linear regressions carried out between the
concentrations of remaining masses and carbon
concentrations were significant in four of the nine
treatments performed (coppice chestnut leaves at the
chestnut orchard (R2 = 0.573, P = 0.018), oak leaves
at the chestnut coppice (R2 = 0.817, P = 0.001),
coppice chestnut leaves in their own environment (R2
= 0.523, P = 0.028), and orchard chestnut leaves in
their own environment (R2 = 0.555, P = 0.021).
Fig. 3 shows the evolution of nitrogen contents in
the decomposing leaves. It was observed that there
were slight initial losses of nitrogen in the treatments
involving oak leaves at the chestnut orchard, orchard
the chestnut leaves in their own environment, orchard
chestnut leaves at the chestnut coppice, oak leaves at
the oak stand, and coppice chestnut leaves at the oak
stand (Fig. 3).
318 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

Table 3 Mean decomposition rates (K0, year-1) [21] and associated parameters (1/k, T50, T95, T99).
Tm T50 T95 T99
Plot Leaves
K0 (year-1) Years Years Years Years
CO 0.44 2.17 1.51 6.52 10.87
C. sativa orchard (CO)
CC 0.46 2.70 1.87 8.11 13.51
Oak 0.37 2.27 1.58 6.82 11.36
CO 0.48 2.17 1.51 6.52 10.87
C. sativa coppice (CC) CC 0.46 3.03 2.10 9.09 15.15
Oak 0.33 2.08 1.44 6.25 10.42
CO 0.50 2.33 1.61 6.98 11.63
Q. pyrenaica (Oak) CC 0.43 2.63 1.82 7.89 13.16
Oak 0.38 2.00 1.39 6.00 10.00

Table 4 Total litter and leaf litter decomposition indexes.

Plots A (kg ha-1) F (kg ha-1) A + F (kg ha-1) K P K0 Kd
Total litter
CO 4,972.20 8,888.97 11,042.66 0.20 420.04 0.24 0.80
CC 5,236.59 9,551.66 14,788.25 0.35 1,854.30 0.55 0.65
Oak 3,492.87 8,226.28 11,719.15 0.30 1,041.04 0.42 0.70
Leaves
CO 2,153.69 2,646.47 4,800.16 0.45 966.30 0.81 0.55
CC 3,101.18 3,743.87 6,845.05 0.45 1,405.00 0.83 0.55
Oak 2,623.21 4,286.16 6,909.37 0.38 995.93 0.61 0.62
A: annual production; F: litter or leaves accumulated in the soil; K: Jennys index; Ko: Olsons index; P: annual loss of produced
fallen litter or leaves; Kd: coefficient of accumulation of fallen litter or leaves. The constants and parameters are according to the
equation K = A/(A+F), P = AK, Ko = A/F, Kd = (AP)/A [23].
Table 5 Correlations between decomposition rates and climatic variables.
Rainfall/Mean T Moisture content
Plot Leaves Rainfall (l m-2) Mean T (C) Max T (C) Min T (C)
(l m-2/C) of the litter %
CO -0.439 0.040 0.021 0.061 -0.387 0.719(**)
C. sativa
CC -0.440 0.045 0.021 0.072 -0.372 0.641(**)
orchard (CO)
Oak -0.453 0.105 0.080 0.134 -0.397 0.652(**)
CO -0.336 -0.005 -0.003 0.020 -0.193 0.587(*)
C. sativa
CC -0.391 0.024 0.038 0.038 -0.254 0.595(*)
coppice (CC)
Oak -0.534(*) 0.198 0.198 0.213 -0.439 0.593(*)
CO -0.347 0.003 0.008 0.029 -0.034 0.607(*)
Q. pyrenaica
CC -0.360 -0.009 0.023 0.006 -0.029 0.587(*)
(Oak)
Oak -0.416 0.070 0.104 0.086 -0.119 0.649(**)
Significant correlations (correlation coefficient of Spearman: * = P < 0.1; ** = P < 0.01).

The linear regressions obtained between the
nitrogen concentration and remaining mass were
significant and negative for five treatments: (oak
leaves at the chestnut orchard (R2 = 0.519, P = 0.002),
orchard chestnut leaves at the at the oak stand (R2 =
0.293, P = 0.030), orchard chestnut leaves at the
chestnut coppice (R2 = 0.561, P = 0.001), coppice
chestnut leaves at the oak stand (R2 = 0.308, P = 0.026)
and oak leaves at the chestnut coppice (R2 = 0.444, P
= 0.005). A similar type of behaviour related to the
initial nitrogen remaining was observed in the leaves
placed in the three plots, but occasional divergences.
During the first three months of the two years, a
noteworthy loss of weight was observed (Fig. 3).
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics 319
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

CO
600,0 CC
Oak
550,0
Carbon (mg g -1 )

500,0

450,0

400,0
0 245 457 698 890
winte spring summer winter spring summer winter spring Days

(a)

CO
600,0
CC
Oak
550,0
Carbon (mg g -1 )

500,0

450,0

400,0
0 245 457 698 890
winter spring summer winter spring summer winter spring Days

(b)

6 0 0 ,0 CO
CC
Carbon (mg g -1 )

5 5 0 ,0 Oa k

5 0 0 ,0

4 5 0 ,0

4 0 0 ,0
0 245 457 698 890
winter spring summer winter spring summer winter spring Days

(c)
Fig. 2 Carbon loss during decomposition of leaves of orchard chestnut (CO), coppice chestnut and oak in the (a) chestnut orchard,
(b) the chestnut coppice, and (c) the oak forest (n = 3).

The two-way analysis of variance-treatment and the case of treatment and interaction, this was not the
year-carried out to determine whether there were case. That is, nitrogen release was greater during the
significant differences in the remaining first year than in the second one in all the treatments
concentrations of nitrogen only revealed significant performed. According to our findings, the release of
differences for the year factor (P = 0.000), while in a key element such as nitrogen in the biogeochemical
320 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

CO
0,20
CC
Nitrgen (g)

0,15 Oak

0,10

0,05

0,00
0 121 245 360 457 586 698 824 Days
winter spring summer winter spring summer winter spring

(a)

CO
0,20 CC
Nitrogen (g)

0,15 Oak

0,10
0,05
0,00
0 121 245 360 457 586 698 824
winter spring summer winter spring summer winter spring Days

(b)

CO
0,20 CC
Oak
Nitrogen (g)

0,15
0,10
0,05
0,00
0 121 245 360 457 586 698 824
winter spring summer winter spring summer winter spring Days

(c)
Fig. 3 Nitrogen loss during decomposition of leaves of orchard chestnut (CO), coppice chestnut and oak in the (a) chestnut orchard,
(b) the chestnut coppice, and (c) the oak forest (n = 3).

cycle is not affected by species, by the substrate on
which the leaves decompose, or by the forest
management regime.
The general trend was for the C/N ratio to decrease
during the course of decomposition, although frequent
increases were detected (Fig. 4). Only in the
treatments involving coppice chestnut leaves at the
oak stand and orchard chestnut leaves at the oak stand
were values higher than the initial ones observed.
Nevertheless, in these two treatments, in the last but
one collection the values of the C/N ratio were similar
to those recorded in other treatments; i.e., around
35-40. The behaviour of the C/N ratio is thus to a
certain extent inverse to that found for nitrogen, since
when there is a high N content, the C/N ratio is low
and vice-versa (Fig. 4).
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics 321
under Different Forest Management in the Sierra de Francia, Salamanca, Spain

75,0
CO
CC
60,0 Oak
C/N

45,0

30,0

15,0
0 245 457 698 890
winter spring summer winter spring summer winter spring Days

(a)

CO
CC
75,0 Oak

60,0

45,0
C/N

30,0

15,0

0,0 Days
0 245 457 698 890
winter spring summer winter spring summer winter spring

(b)

75,0
CO
CC
60,0 Oak
C /N

45,0

30,0

15,0
Days
0 245 457 698 890
winter spring summer winter spring summer winter spring

(c)
Fig. 4 C/N ratio during decomposition of leaves of orchard chestnut (CO), coppice chestnut and oak in the (a) chestnut orchard, (b)
the chestnut coppice, and (c) the oak forest (n = 3).
322 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
The linear regressions obtained between the
remaining mass and the C/N ratio were significant
for three treatments (the orchard chestnut leaves in
their own environment (P = 0.028), the coppice
chestnut leaves in their own environment (P = 0.001),
and the orchard chestnut leaves at the chestnut
coppice (P = 0.016)). The control of the C/N ratio
over decomposition in our study is thus not very
strong.
The two-way analysis of variance-treatment and
month-for the C/N ratio only revealed significant
differences for the year factor (P = 0.001).
4. Discussion
4.1 Changes in Dry Mass and Decomposition Rates
In all cases, we observed a general trend towards a
gradual decrease in weight, a two-phase model being
observed (Fig. 1). During the first two months of the
nine treatments, a marked weight loss was observed;
thereafter, and until the end of the process, a weaker
phase of weight loss was seen, characterised by both
increases and losses. The high degree of initial
mineralization may be due to the leaching of soluble
materials contained in the leaf litter and consumption
by microorganisms of the labile material of the leaf
litter [21, 24].
This first phase is governed by continuous
mineralization, and the loss in the three plots is very
similar, around the 15% (Fig. 1), and by the
concentration of readily leachable soluble components
and of microbial activity-limiting nutrients, such as
nitrogen, phosphorus, sulphur, this has been observed
by many authors [25-28].
The second phase of the decomposition process in
the dry summer period (Fig. 1) was characterised by a
saw-toothed decrease in the remaining mass. In the
three plots mineralization slowing down when the leaf
litter is dry. This is a period with little moisture in the
topsoil, while throughout the autumn and winter,
coinciding with the rainy season, the activity of
decomposer microbiota was favoured, and more
substantial losses of material occurred [29].
Throughout this second phase, although a clear trend
towards weight loss was observed, some increases in
mass were observed in some of the samplings (Fig. 1).
These increases in weight may have been due to
contamination by dust, increases in the heterotrophic
mass associated with the leaf litter, and the random
conditions to which the samples were subjected owing
to the variability of the environmental conditions, the
chemical composition of the leaf litter, and the
microbiota responsible for the degradation of plant
remains [23]. In contrast to the first phase of the
decomposition process, in this second phase the
process is governed by the more complex and
decomposition resistant organic matter [30-32].
In the present study, neither species nor the
different forest management regimes to which the
chestnuts were subjected seemed to affect the mass
loss dynamics of the leaves in the litterbags, such that
it seems that decomposition must be governed by
other factors. The fact that we failed to observed
differences in the process of mass loss among the
treatments performed could have been due to the
similarity of the climatic conditions prevailing at all
three sites and the similarity in the chemical
composition and physical factors of the leaves, such as
hardness or texture.
The highest rates corresponded to the first months,
when the losses of dry mass were more intense; later,
as the process advanced, the decomposition rates were
lower, owing to the smaller amount of dry matter
mineralized (Table 2).
These decomposition rates lie within the general
range for estimations of k. The F values may be
underestimated, since it is often difficult to distinguish
decomposing leaves from other plant remains. In a
compilation of the bibliographic data, Santa Regina
[33] reported a value of 0.35 year-1 for the degradation
constant k in the Mediterranean region.
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
The time required for 50% of the initial mass of the
material to decompose is known as the half-life and is
equivalent to 0.693/k. This is a very useful descriptive
parameter for comparing the duration of leaf litter
decomposition, even in cases in which the exponential
model is not applicable [22].
The oak leaves in their own environment showed
the lowest half-life values, for that, it seems this was
the leaf litter that would decompose fastest. Intead, the
coppice chestnut leaves in the three plots showed the
highest half-life values; i.e., they would be the leaves
that decompose more slowly. Aerts [34], in different
oak species, have obtained different equations that
indicated half-decomposition times (50% of the initial
matter) of 32 months for a group of Q. pyrenaica plots
and 26 months for a second group. The results in the
literature are sometimes conflictive since they are
based on both field and laboratory studies. Rapp et al.
[35] obtained a decomposition rate of 50% for 25
months in Q. ellipsoidalis, while Bockheim et al. [36],
under controlled moisture conditions, recorded
half-decomposition times in Q. ilex, Q. coccifera and
Q. lanuginosa leaves after 22 months of
decomposition.
The highest values were found during the first
months, coinciding with the most marked mass losses.
During the other months, the monthly rates were
considerably lower, and on some occasions even
negative rates were obtained (Table 2). These are the
result of the way in which the decomposition was
assessed, based on the weights of different samples
taken over successive months. Thus, perhaps
accidentally, it is possible that a sample taken in a
given month could weigh more than in the previous
one. As mentioned when describing the loss of dry
mass, these increases are fairly frequent in the
measurements of leaf litter decomposition using the
method of Rapp [37], above all in the later stages of
the process. In the present study, negative annual rates
were more frequent in the summer months, the period
in which owing to the prevalence of high temperatures
323
and low soil moisture contents, the process slows
down. During the autumn months, which are much
more favourable for the degradation of plant matter, it
was less frequent to observe negative monthly rates.
The negative simple exponential model has been
widely used [20, 22, 27, 29, 38-41]. Although other
models can be used to describe the decomposition
process, the negative simple exponential model is
most widely used, since it is able to describe
decomposition closely and mass loss is characterized
by a single constant, thereby facilitating comparisons
with the results of other studies [42-44].
In leaves with high contents of both nitrogen and
lignin decomposition may be slowed or may even halt
completely, because such compounds react with one
another, giving rise to recalcitrant compounds that are
not readily degraded by soil microorganisms [45].
4.2 Decomposition Indices
We hardly observed any differences between the
decomposition rates calculated from the leaf litterbag
model and those obtained from the material
accumulated at ground level and the annual
contributions (Table 4). In general, the decomposition
constants deduced from the leaf litterbag model were
higher. This can be explained that the bags only
included leaves, whereas the ground also included
ligneous material, of slower decomposition, together
with partially decomposed organic remains that had
already lost their labile components, such that their
degradation was also fairly slow [46, 47]. Moreover,
the environment inside the litter leaf bags is somewhat
different from that of the exterior, since better
substrate homogeneity is achieved as well as greater
moisture in the leaves, both factors favouring the
probability that an overestimation of the true
decomposition would occur [23].
Despite the foregoing, in the literature it is also
possible to find cases in which the decomposition
constants calculated with the litterbag method are
higher [48]. According to Tarazona et al. [49], the
324 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
mesofauna may encounter difficulty in accessing the
inside of the bags, such that the fragmentation and
consumption of leaves decrease and, additionally, the
microclimatic conditions inside the bags are less
favourable for decomposition.
4.3 Control of Decomposition by the Climate
Regarding the effects of temperature, no correlation
between this factor and the decomposition rate was
observed in any of the three litter types (treatments)
used across three different sites (Table 5). In the
Mediterranean climate, temperatures never reach
sufficiently high levels for microbial activity to be
inhibited except on a few very special occasions
throughout the year; rather they are quite gentle and
favourable for the activity of the decomposing
microbiota. When a negative effect on decomposition
is observed, this is due to the coincidence in time of
high temperatures and periods of drought [50].
In Mediterranean climates, characterized by an
alternation between wet and dry seasons, the influence
of moisture on decomposition is fairly strong [51].
These humectation and desiccation processes affect
the growth of the microbial population to a
considerable extent, thus affecting the flows of C and
N mineralization [52]. Owing to the high temperatures
prevailing in summer the soil surface is dried, which
prevents the establishment of a stable microflora on
the leaf litter, and this in turn leads to a shift in the
conditions ideal for mineralization towards deeper
horizons, where moisture tends to remain constant
[53].
4.4 Carbon and Nitrogen Dynamics in Decomposing
Leaves
Overall, the evolution of organic carbon during the
decomposition process courses parallel to the loss of
mass (Fig. 2), since the structural components of the
leaves mainly contain carbon [54, 55].
Two-way analyses of variance were
performed-treatment and year, but these only revealed
significant differences for year coppice chestnut
leaves at the chestnut orchard (P = 0.000), with no
differences for either treatment or the interaction. That
is, the release of carbon from the decomposing leaves
did not differ among the nine treatments studied, but
there were differences between the two study years, in
the sense that a greater release of carbon was observed
in the first year than in the second one owing to the
strong initial losses of water-soluble material from the
leaves.
At the end of the 2.4 years of the study, most of the
treatment assays higher nitrogen concentrations than
the initial ones were recorded. Concentrations lower
than the initial ones were only seen in the treatments
with orchard chestnut leaves at the oak stand and oak
leaves at the chestnut orchard, and concentrations
equal to the initial ones were seen in the treatment
with coppice chestnut leaves at the oak stand.
However, in these three treatments the samplings prior
to the last one revealed nitrogen concentrations higher
than the initial ones, which may be concluded that the
trend was towards an increase in concentration.
The first phase in the decrease in nitrogen
concentrations has been addressed by other authors [4,
56, 57] and has been attributed to the washing out of
labile nitrogen components [58]. The duration of this
varied between two and four months (Fig. 3);
thereafter, nitrogen concentrations increased. In the
rest of the treatments, this increase in nitrogen
concentrations was observed from the first sampling.
Increases in nitrogen concentrations have often been
observed in leaf litter decomposition [26, 34, 39, 56,
59, 60] and can be attributed to several causes. The
main one is microbial immobilization. The duration of
the accumulation phase may vary between a few
months and several years [10, 61, 62]. In our case, the
two and a half years of the experimental period might
not have been sufficient to gain an understanding of
the whole process.
Most authors have used the C/N ratio to determine
the moment at which nitrogen ceases are to be
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
determinant for the microbiota and begin to be
mineralized. There is a critical C/N ratio above since
nitrogen is immobilized. When the C/N ratio is high,
this means that the amount of nitrogen is low and
hence decomposer microorganisms immobilize it in
their tissues. As decomposition advances, the carbon
is mineralized, such that the C/N ratio declines
progressively, and when it reaches a level below the
critical value the release of nitrogen begins in a way
proportional to the loss of mass [4]. The C/N ratio
indicates the relationship between nitrogen dynamics
and the loss of mass during decomposition. Thus, the
ratio increases nitrogen is released through
mineralization or leaching. If the ratio remains
constant, nitrogen release follows a model similar to
that of mass. If the C/N ratio decreases, nitrogen is
accumulated in biological, physical or chemical form.
Santa Regina [23] established a value of 30 for the
critical C/N ratio above which nitrogen is immobilized
and below which it is mineralized. In the present work,
the values of the C/N ratio were between 59.16 in the
treatment involving orchard chestnut leaves at the oak
stand and 26.81 in the treatment with orchard chestnut
leaves in their own environment (Fig. 4). Since
nitrogen continues to be immobilized at the end of the
treatment, these values could be too high, the
threshold C/N ratio will be smaller, and hence at the
end of the study period the process continues in the
immobilization stage.
5. Conclusions
Regarding the decomposition of leaf litter, in all the
treatments performed here the weight loss of dry
matter followed a biphasic pattern that included an
initial stage controlled by readily leachable materials
and characterized by strong losses during the first
months. This was followed by a second, saw-toothed
decrease in the remaining mass, governed by organic
matter that was more complex and more resistant to
decomposition. Accordingly, the decomposition
dynamics of the coppice and orchard chestnuts were
325
similar, and no differences were observed with respect
to the dynamics observed at the oak stand,
representing the climax vegetation of the area.
The oak leaves had the lowest half-life values in
their own environment; that is, this was the type of
leaf litter that would decompose fastest. In contrast,
the coppice chestnut leaves in their own environment
showed the highest half-life values; i.e., they would be
the leaves that decompose more slowly. Coppice
chestnut leaves degraded more slowly not only at their
own plot but also at the other two plots.
The parametric two-way analysis of variance of the
loss of dry mass, carbon and nitrogen, only detected
significant differences for the year factor, with no
differences for either treatment or the interaction. This
confirms that the remaining masses of decomposing
leaves were significantly larger in the first year with
respect to the second one in the nine treatments, but
no differences in mass loss were observed between
them throughout the study.
In the experiments studied, carbon and nitrogen
from the decomposing leaves showed a peculiar type
of behaviour, with the finding of elements that had
higher concentrations at the end of the process than at
the beginning. Increases in the concentration of carbon
and nitrogen, both absolute and relative, have been
observed. Microbial fixation of atmospheric nitrogen
contributes to this increase since there is an abundant
source of carbon energy in the leaf litter and suitable
humidity and temperature for nitrogen fixers.
References
[1] F. Pardo, L. Gil, J.A. Pardos, Field study of beech (Fagus
sylvatica L.) and oak (Quercus pyrenaica Will.) leaf litter
decomposition in the centre of the Iberian Peninsula,
Plant and Soil 191 (1997) 89-100.
[2] V. Meentemeyer, Macroclimate and lignin control of
litter decomposition rates, Ecology 59 (1978) 465-472.
[3] B. Berg, H. Staaf, Decomposition rate and chemical
changes of Scot pine needle litter: II. Influence of
chemical composition, in: Structure and Function of
Northern Coniferous Forests, An Ecosystem Study,
Ecological Bulletins (Stockholm), vol. 32, 1980, pp.
373-390.
326 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
[4] B. Berg, H. Staaf, Leaching, accumulation and release of
nitrogen in decomposition forest litter, Ecological
Bulletins 33 (1981) 163-178.
[5] J.M. Melillo, J.D. Aber, J.F. Muratore, Nitrogen and
lignin control of hardwood leaf litter decomposition
dynamics, Ecology 63 (1982) 621-626.
[6] T. Sariyildiz, J.M. Anderson, Variation in the chemical
composition of green leaves and leaf litters from three
deciduous tree species growing on different soil types,
Forest Ecology and Management 210 (1-3) (2005)
303-319.
[7] T. Sariyildiz, J.M. Anderson, M. Kucuk, Effects of tree
species and topography on soil chemistry, litter quality
and decomposition in Northeast Turkey, Soil Biology and
Biochemistry 37 (9) (2005) 1695-1706.
[8] M. Jacob, N. Weland, C. Platner, S.M. Chaefer, C.
Leuschner, F.M. Thomas, Nutrient release from
decomposing leaf litter of temperate deciduous forest
trees along a gradient of increasing tree species diversity,
Soil Biology and Biochemistry 41 (10) (2009)
2122-2130.
[9] T.R. Seastedt, The role of microarthropods in
decomposition and mineralization processes, Annual
Review of Entomology 29 (1984) 25-46.
[10] A. Fioreto, S. Papa, A. Pellegrino, A. Fuggi,
Decomposition dynamics of Myrtus communis and
Quercus ilex litter: Mass loss, microbial activity and
quality change, Applied Soil Ecology 36 (2007) 32-40.
[11] M.B. Johansson, B. Berg, W. Meentemeyer, Litter
mass-loss rates in late stages of decomposition in a
climatic transect of pine forest, long-term decomposition
in a Scots pine forest, Canadian Journal of Botany 73
(1995) 1509-1521.
[12] S. Salazar, I. Santa-Regina, Organic matter and nutrient
dynamics of the litter layer on a two chestnut stands in
Honfra area, Spain, ISHS section nuts and Nediterranean
climate fruits, Acta Horticulturae 693 (2005) 691-700.
[13] M.P. Davey, B. Berg, B.A. Emmett, P. Rowland,
Decomposition of oak leaf litter is related to initial litter
Mn concentrations, Canadian Journal of Botany 85 (1)
(2007) 16-24.
[14] F. Raimundo, A. Martins, M. Madeira, Decomposition of
chestnut litterfall and eight-year soil chemical changes
under a no-tillage management system in Northern
Portugal, Annals of Forest Science 65 (4) (2008)
408-415.
[15] T. Sariyildiz, Litter decomposition of Picea orientalis,
Pinus sylvestris and Castanea sativa trees grown in
Artvin in relation to their initial litter quality variables,
Turkey Journal of Agriculture and Forestry 27 (2003)
237-243.
[16] T. Sariyildiz, J.M. Anderson, Decomposition of sun and
shade leaves from three deciduous tree species: As
affected by their chemical composition, Biology and
Fertility of Soils 37 (2003) 137-146.
[17] M. Ilangovan, K. Paliwal, Changes in mass and nutrients
during decomposition of Leucaena leucocephala and
Cymbopogon citrates and the effect of substrate quality,
weather variables on mass loss during decompostion in a
semiarid ecosystem, Maduari, India, Journal of Tropical
Forest Science 8 (1996) 317-332.
[18] T. Sariyildiz, M. Kucuk, Litter mass loss rates in
deciduous and coniferous trees in Artvin, Northeast
Turkey: Relationship with litter quality, microclimate and
soil characteristics, Turkish Journal of Agriculture and
Forestry 32 (6) (2008) 547-559.
[19] T.R. Moore, J.A. Trofymow, B. Taylor, C. Prescoo, C.
Camire, L. Duschene, et al., Litter decomposition rates in
Canadian forest, Global Change Biology 5 (1999) 75-82.
[20] S. Salazar, Estudio de procesos ecolgicos para el
desarrollo sostenible del castao (Castanea sativa Mill.)
de la Sierra de Francia, Tesis Doctoral, Universidad de
Salamanca, 2008, p. 327.
[21] J.S. Olson, Energy storage and the balance of producers
and decomposers in ecological systems, Ecology 14
(1963) 322-331.
[22] H. Enny, S.P. Gessel, F.T. Binghan, Comparative study
of decomposition rates of organic matter in temperate and
tropical regions, Soil Science 68 (1949) 419-432.
[23] I. Santa Regina, Litter fall, decomposition and nutrient
release in three semi-arid forest of the Duero Basin, Spain,
Forestry 74 (4) (2001) 347-358.
[24] A. Smolander, V. Kitunen, L. Paavolainen, E. Mlknen,
Decomposition of Norway spruce and Scots pine needles:
Effects of liming, Plant and Soil 179 (1996) 1-7.
[25] A. Martn, J.F. Gallardo, I. Santa-Regina, Long-term
decomposition process of the leaf litter from Quercus
pyrenaica forest across a rainfall gradient (Spanish
central system), Annals of Forest Science 54 (1997)
91-202.
[26] E.A. Webster, C. Halpin, J.A. Chudek, E.L. Tilston, D.W.
Hopkins, Decomposition in soil of soluble, insoluble and
lignin-rich fractions of plant material from tobacco with
genetic modifications to lignin biosynthesis, Soil Biology
and Biochemistry 37 (2005) 751-760.
[27] J. Huang, X. Wang, E. Yan, Leaf nutrient concentration,
nutrient resorption and litter decomposition in an
evergreen broad-leaved forest in eastern China, Forest
Ecology and Management 239 (2007) 150-158.
[28] A. Martnez-Yrzar, S. Nez, A. Brquez, Leaf litter
decomposition in a southern Sonora desert ecosystem,
Northwestern Mexico: Effects of habitat and litter quality,
Acta Oecologica 32 (2007) 291-300.
[29] S. Imgraven, S. Dittmann, Leaf litter dynamics and litter
 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
consumption in two temperate South Australian
mangrove forest, Journal of Sea Research 59 (2008)
83-93.
[30] J.D. Lousier, D. Parkinson, Litter decomposition in a cool
temperate deciduous forest, Canadian Journal of Botany
54 (1976) 419-436.
[31] M. Rapp, S. Leonardi, Evolution de la litire au sol au
cours dune anne dans un taillis de chne vert (Quercus
ilex), Pedobiology 32 (1988) 177-185.
[32] G. Almendros, J. Dorado, F.J. Gonzlez-Vila, M.J.
Blanco, U. Lances, C-13 NMR assessment of
decomposition patterns during composting of forest and
shrubs biomass, Soil Biology and Biochemistry 32 (6)
(2000) 793-804.
[33] I. Santa Regina, Organic matter distribution and nutrient
fluxes within a sweet chestnut (Castanea sativa Mill.)
stand of the Sierra de Gata, Spain, Annals of Forest
Science 57 (2000) 691-700.
[34] R. Aerts, Climate, leaf litter chemistry and leaf litter
decomposition in terrestrial ecosystems: A triangular
relationship, Oikos 79 (1997) 439-449.
[35] I. Santa Regina, M. Rapp, A. Martn, J.F. Gallardo,
Nutrient release dynamics in decomposing leaf litter in
two Mediterranean deciduous oak species, Annals of
Forest Science 54 (1997) 747-760.
[36] J.G. Bockheim, E.A. Jensen, D.M. Heisey, Nutrient
dynamics in decomposing leaf litter of four tree species
on a sandy soil in North Western Wisconsin, Canadian
Journal of Forest Research 21 (1991) 803-812.
[37] M. Rapp, Etude exprimentale de la libration dlments
minraux lors de la dcomposition de litires dessences [51]
mditerranennes, CR Acad. Sc. Paris 264 (1967) 797.
[38] K.L. Bocock, O.J. Gilbert, The disappearance of leaf
litter under different woodland conditions, Plant and Soil
9 (1957) 179-185.
[39] E.F. Salamanca, N. Kaneko, S. Katagiri, Rainfall
manipulation effects on litter decomposition and the
microbial biomass of the forest floor, Applied Soil
Ecology 22 (2003) 271-281.
[40] A. Fioretto, C. Di Nardo, S. Papa, A. Fuggi, Lignin and
cellulose degradation and nutrient dynamics during
decomposition of three leaf litter species in a
Mediterranean ecosystem, Soil Biology and Biochemistry
1 (2005) 1083-1091.
[41] A.R. Barbhuiya, A. Auranachalam, P.C. Nath, L. Khan,
K. Aurachalam, Leaf litter decomposition of dominant
tree species of Namdapha National Park, Arunachal
Pradesh, Northeast India, Journal of Forest Research 13
(2008) 25-34.
[42] Q. Wang, S. Wang, Y. Huang, Comparisons of litterfall,
litter decomposition and nutrient return in a monoculture
Cunninghamia lanceolata and mixed stand in southern
327
China, Forest Ecology and Management 255 (2008)
1210-1218.
[43] R.K. Wieder, G.E. Lang, A critique of the analytical
methods used in examining decomposition data obtained
from litter bags, Ecology 63 (1982) 1636-1642.
[44] T. Sariyildiz, A. Tfekciogluve, M. Kck, Comparison
of decomposition rates of beech (Fagus orientalis Lipsky)
and Spruce (Picea orientalis (L.) Link) litter in pure and
mixed stands of both species in Artvin, Turkey, Turkish
Journal of Agriculture and Forestry 29 (2005) 429-438.
[45] T. Sariyildiz, Effects of gap size on long-term litter
decomposition rates of beech, oak and chestnut species in
a high-elevation of Northeast Turkey, Ecosystems 11 (6)
(2008) 841-853.
[46] K. Fog, The effect of added nitrogen on the
decomposition of organic matter, Biological Reviews 63
(1988) 433-462.
[47] P.M. Attiwill, M.A. Adams, Nutrient cycling in forest,
New Phytologist 124 (1993) 561-582.
[48] I. Santa Regina, J.F. Gallardo, Biogeochemical cycles in
forests of the Sierra de Bjar mountains (province of
Salamanca, Spain): Decomposition index of the leaf litter,
Annals of Forest Science 52 (1995) 393-399.
[49] I. Santa Regina, T. Tarazona, Dynamics of litter
decomposition in two Mediterranean climatic zone forest
of the Sierra de la Demanda, Burgos, Spain, Arid Soil
Research and Rehabilitation 9 (1995) 201-207.
[50] R.G. Joergensen, Organic matter and nutrient dynamics
of the litter layer on a forest Rendzina under beech,
Biology and Fertility of Soils 11 (1991) 163-169.
M.V. Garrido, Ciclos de nutrientes: Produccin,
descomposicin y dinmica de los elementos minerales
de los detritos arbreos, Tesis Doctoral, Universidad de
Salamanca, 2001, p. 251.
[52] M.M. Coteaux, P. Bottner, B. Berg, Litter
decomposition, climate and litter quality, Trends in
Ecology and Evolution 10 (1995) 63-66.
[53] P. Bottner, Response of microbial biomass to alternate
moist and dry conditions in a soil incubated with 14C and
15N-labelled plant material, Soil Biology and
Biochemistry 7 (1985) 329-337.
[54] P. Rovira, V.R. Vallejo, Organic carbon and nitrogen
mineralization under Mediterranean climatic conditions:
The effects of incubation depth, Soil Biology and
Biochemistry 29 (9-10) (1997) 1509-1520.
[55] J.F. Gallardo, I. Santa-Regina, A.F. Harrison, M. Howard,
Organic matter and nutrient dynamics in three ecosystems
of the Sierra de Bejar mountains (Salamanca province,
Spain), Acta Oecologica 16 (4) (1995) 447-459.
[56] E. Garca, Efecto sobre el suelo de la hojarasca de
arbustos en la Sierra de Guadarrama, Tesis Doctoral,
Universidad Complutense de Madrid, 2001, p. 259.
328 Long-term Decomposition Process of the Leaf Litter, Carbon and Nitrogen Dynamics
under Different Forest Management in the Sierra de Francia, Salamanca, Spain
[57] J.A. Blanco, La prctica de las claras forestales y su Japanese cedar (Cryptomeria japonica D. Don) plantation,
influencia en el ciclo interno de nutrientes en dos bosques Forest Ecology and Management 234 (2006) 192-198.
de pino silvestre de los Pirineos Navarros, Tesis Doctoral, [61] R.R. Pandey, G. Sharma, S.K. Tripathi, A.K. Singh,
Universidad Pblica de Navarra, 2004, p. 365. Litterfall, litter decomposition and nutrient dynamics in a
[58] T, Miyamoto, T. Hiura, Decomposition and nitrogen subtropical natural oak forest and managed plantation in
release from the foliage litter of fir (Abies sachalinensis) Northeastern India, Forest Ecology and Management 240
and oak (Quercus crispula) under different forest (2007) 96-104.
canopies in Hokkaido, Japanese of Ecology Research 23 [62] D. Albers, S. Migge, M. Schaefer, S. Scheu,
(4) (2008) 673-680. Decomposition of beech leaves (Fagus sylvatica) and
[59] A.T. Austin, P.M. Vitousek, Precipitation, decomposition spruce needles (Picea abies) in pure and mixed stands of
and litter decomposability of Metrosideros polymorpha in beech and spruce, Soil Biology and Biochemistry 36
native forest on Hawai, Journal of Ecology 88 (2000) (2004) 155-164.
129-138. [63] S.R. Isaac, M. Nair, Biodegradation of leaf litter in the
[60] S.J. Joo, M.Y. Yim, K. Nakane, Contribution of warm humid tropics of Kerala, India, Soil Biology and
microarthropods to the decomposition of needle litter in a Biochemistry 37 (2005) 1656-1664.
Journal of Agricultural Science and Technology B 2 (2012) 329-338
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Linking Markets to Smallholder Agro-forestry Farmers

as a Strategy for Poverty Alleviation in the Tropics

Twaha Ali Basamba1, Barnabas Kiiza2, Clement Mayanja2, Bob Nakileza3, Frank Matsiko4, Paul Nyende5 and
Elizabeth Bacwayo Kukunda6
1. Department of Agricultural Production, College of Agricultural and Environmental Sciences, Makerere University, P.O. Box 7062,
Kampala, Uganda
2. Department of Agribusiness and Natural Resource Economics, College of Agricultural and Environmental Sciences, Makerere
University, P.O. Box 7062, Kampala, Uganda
3. Department of Geography, Geo-informatics and Climatic Sciences, College of Agricultural and Environmental Sciences, Makerere
University, P.O. Box 7062, Kampala, Uganda
4. Department of Extension and Innovation Studies, College of Agricultural and Environmental Sciences, Makerere University, P.O.
Box 7062, Kampala, Uganda
5. Africa 2000 Network Uganda, P.O. Box 21990, Kampala, Uganda
6. Department of Development Studies, Uganda Christian University, P.O. Box 4, Mukono, Uganda

Received: August 31, 2011 / Published: March 20, 2012.

Abstract: The objective of the study was to assess factors affecting market participation of agro-forestry smallholder farmers in
Uganda. Primary data on household, farm and marketing characteristics were collected from 153 farming households using a
semi-structured, pre-tested questionnaire. The determinants of market participation were assessed using the Probit model. Descriptive
statistics showed that a combination of crop, livestock and tree products were marketed by farmers. The main products included
maize, cassava and coffee (crops); firewood and poles (tree products); and birds and goats (livestock products). Results from the
Probit model showed that farm size, household size, education level, access to credit and extension visits had positive and statistically
significant effects on market participation of farmers in agro-forestry products markets. Age of farmer had a negative and
statistically significant effect on the participation. Emphasis on improving the quality and coverage of extension services, extending
credit facilities to farmers and intensifying agro-forestry training among farmers are suggested as avenues to enhance participation of
farmers in agro-forestry products markets.

Key words: Agro-forestry, smallholder farmers, poverty alleviation, market participation, tropics.
.

1. Introduction combination with crops and livestock is an ancient

practice. However, several factors have contributed to
Agro-forestry is a dynamic, ecologically based
a rising interest in agro-forestry since the 1970s,
natural resources management system that, through
including the deteriorating economic situation in
the integration of trees on farms and in the agricultural
many parts of the developing world, increased tropical
landscape, diversifies and sustains production for
deforestation, degradation and scarcity of land due to
increased social, economic and environmental benefits
population pressure and the growing interest in
for land users at all levels [1]. Cultivating trees in
farming systems, intercropping and the environment
[2]. Brandle et al. [3], Clason [4] and Kurtz [5] argued
Corresponding author: Twaha Ali Basamba, Ph.D., senior
lecturer, research fields: agro-forestry and land management. that agro-forestry provides cost-effective alternatives
E-mail: twaha@agric.mak.ac.ug.
330 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
that increase farm profits and protects the environment
within rural settings. In such a scenario, agro-forestry
practices that integrate tree, crop and animal
components are emerging as part of an intensive land
management approach focused on sustainable
resource use and production within given economic,
social and environmental settings [6, 7]. The
ecological, economic and social benefits of
agro-forestry are further documented [8-11]. With its
potential to increase farm productivity and diversity,
agro-forestry produces a range of products such as
food, fuelwood, building materials, medicine and
fodder. Properly conceived and practiced,
agro-forestry can contribute to the sustained
productivity of the natural resource base by enhancing
soil fertility, controlling soil erosion, enhancing the
microclimate of cropping and grazing lands and
generally improving the environment. Agro-forestry
has the potential to reduce poverty and can efficiently
be used in poverty reduction strategies of the tropical
East African countries. In forest-scarce African
countries, agro-forestry has expanded greatly on small
farms. In Kenya and Ethiopia, for example,
agro-forestry farms account for most of the timber and
pole production [12]. The Ugandan governments
long-term plan for modernization of agriculture [13]
also recognizes smallholder agro-forestry as a strategy
for poverty alleviation and ecosystem protection in the
country. In the plan for modernization of agriculture,
poverty is defined as a situation when one or more
persons fail to attain a level of well-being (usually
material) that is deemed to constitute a reasonable
minimum by the standards of the Ugandan society.
However, linking markets to smallholder agro-forestry
farmers still remains a big socio-economic
development challenge in this part of the world.
Markets are of fundamental importance in the
livelihood strategy of most rural households, rich and
poor alike. Markets are where farmers, as producers,
buy their inputs and sell their products; and where, as
consumers, spend their income from the sale of crops
and non-agricultural activities, to buy their food
requirements and other consumption goods. IFAD [14]
argued that rural producers in many parts of the world
cannot improve their living standards because of
serious difficulties in accessing markets. Linking
smallholder farmers to markets has become more
important over time because of the long-term effects
of liberalization of agricultural markets throughout the
developing world [15]. Liberalization policies of
elimination of commodity marketing boards,
relaxation of price controls, privatization of
state-owned companies and elimination of restrictions
on agricultural production and marketing have
increased competition and given farmers greater
responsibility over their production and marketing
decisions. While markets are critical to smallholder
farmers, the factors that influence how markets
operate, such as access to information and market
institutions, have a powerful influence on market
participation and earnings of smallholder farmers.
Little attention has been given to marketing systems in
the smallholders rural livelihoods frameworks and
their role in poverty alleviation. Thus, linking
smallholder agro-forestry farmers to particular niche
markets is critical to the success of agro-forestry
innovations and economic prosperity of these
poverty-ridden peasants in the tropics.
Trees are useful in livelihood and production
strategies, especially among rural communities. Muok
et al. [16] noted that growing trees on farms is a very
important livelihood strategy in rural communities of
sub-Saharan Africa. Agro-forestry, as a science and
practice, has the potential to contribute to the
improvement of rural livelihoods due to the capacity
of its various forms to offer multiple alternatives and
opportunities to smallholders to enhance farm
production and income, while protecting the
agricultural environment. While agro-forestry was
identified as an important production and livelihood
strategy in rural communities of sub-Saharan Africa
[17], there is little consideration for the markets [18],
 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
yet expanding market opportunities for smallholder
farmers are critical to the success of agro-forestry
innovations. A variety of agro-forestry technologies is
finding enormous application in the East and Central
African region. Such technologies are lifting many
smallholder farmers out of poverty and mitigating
declining agricultural productivity and the quality and
quantity of natural resources. While thousands of
smallholder farmers are adopting these technologies,
the key challenge remains lack of access to markets.
Most farmers have little access to market information
concerning the demand and price of agro-forestry
products, little knowledge of market specifications
and weak linkages with market agents. As a result,
they have little knowledge on how to assess the value
of their products and where to market them. In
addition, the review of existing literature and
experiences from past agro-forestry initiatives reveals
little information about the determinants of farmers
participation in agro-forestry markets in Uganda. For
example, Byabashaija et al. [19] studied the traditional
uses of indigenous tree species; Kiwuso et al. [20]
identified the indigenous methods of controlling
termites in Ugandas agro-forestry systems; Okullo et
al. [21] described the local knowledge of Shea butter
tree production; Okorio et al. [22] studied the potential
of agro-forestry in improving livelihoods in Eastern
and mid-Northern Uganda; and Katumba et al. [23]
examined the domestication of medicinal tree species
in the Lake Victoria shore region. These studies reveal
no or limited information on the factors that influence
farmers participation in agro-forestry markets in
Uganda. It is against this background that the study
seeks to fill this gap.
The general objective of this study was to evaluate
the factors affecting market participation of
agro-forestry smallholder farmers. The specific
objectives of the study were: (1) to compare the
socio-economic characteristics of agro-forestry
farmers participating and not participating in markets;
(2) to characterize agro-forestry products marketed by
331
farmers; and (3) to analyze factors influencing
farmers participation in agro-forestry markets. The
research hypotheses of this study were: (1) access to
credit significantly affects farmers participation in
agro-forestry markets; and (2) membership to farmer
organizations increases the probability of participating
in agro-forestry markets.
2. Materials and Methods
2.1 Description of the Study Area
The survey was conducted in Busia District which
is located in the south-eastern part of Uganda, north of
Lake Victoria and west of the Republic of Kenya in
the African tropics. It is approximately 196 km far
from Kampala, the Capital city of the Republic of
Uganda. The district which covers a total area of 743
sq. km accommodates a population of 243,298 people
(117,564 males and 125,734 females) [24]. The
district was selected because tree growing is the third
major source of income for the majority of the farmers,
after crops and livestock [25]. Therefore, agro-forestry
has a great potential to improve farmers livelihoods
[22]. The district is predominantly rural, with 84% of
the population living in rural areas and about 85% of
this population survives on agriculture. The rural
economy of the district is characterized by
smallholder farmers who predominantly practice
subsistence agriculture with the main crops grown
being sorghum, millet, cotton, cassava, sweet potatoes,
maize and beans.
2.2 Sample Size and Sampling Procedures
Using purposive and simple random sampling
techniques, cross-sectional data were collected from a
total sample size of 153 farmers. Dabani and Bulumbi
Sub-Counties in Busia District were purposively
selected to represent high and low levels of adoption
of agro-forestry technologies, respectively. This was
followed by the purposive selection of Nangwe Parish
from Dabani Sub-County and Bubango Parish from
Bulumbi Sub-County where previous research and
332 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
dissemination activities by the government of Uganda,
through the NARO (National Agricultural Research
Organization), had taken place. From each parish, a
list of farmers that was used as a sampling frame was
generated with the help of the district agricultural
extension officers. A simple random sampling
technique was used to select 60 farmers from Nangwe
parish and 93 farmers from Bubango parish to make a
total sample size of 153 farmers for the study.
2.3 Data Collection
Both primary and secondary data were collected
during the study. A semi-structured, pre-tested
questionnaire was used to capture primary data from
selected farmers at their respective farms using direct
interviews. Primary data included farmers age,
gender, marital status, main occupation and farming
experience. It also included household size, number of
household members involved and not involved in
farming, farm ownership, farm size, education level of
household head, education level of spouse,
membership to farmers organizations, access to credit
facilities, access to agro-forestry output markets, main
agro-forestry production objective, contact with
extension personnel, distance to trading centre,
distance to district capital, types and prices of
agro-forestry products. Secondary data included
reviewing of previous research findings on the
marketing of agro-forestry products among
smallholder farmers.
2.4 Analytical Procedures
2.4.1 Probit Model
In order to analyze factors influencing farmers
participation in agro-forestry products markets, the
Probit model was used. A probit model is a
specification for an ordinal or a binary response model
which employs a probit function. The regression
analysis involves a categorical dependent variable,
usually 1 and 0, using a cumulative normal
distribution function. The probit function is the
inverse CDF (Cumulative Distribution Function) or
quantile function associated with the standard normal
distribution. The probit model is estimated using
standard maximum likelihood procedures and such an
estimation is called a probit regression. The
probability that a household participates in a given
activity is a function of the independent (exogenous)
variables and is estimated using the Probit model
because of the advantage, it has over many other
models in that the probabilities are bounded between 0
and 1 [26]. At the same time, it compels the error
terms to be homoscedastic for the reason that the
forms of probability functions depend on the
distribution of the difference between the error terms
associated with one particular type and another. The
Probit model indicates that the observable dependent
variable, yi, is determined as follows, given the latent
(unobservable) random variable y*i [27].
yi = 1 if y*i >0
yi = 0 if y*i <0 (1)
* * *
Given that y i = X i + u i, then the probability that
yi = 1 is given as
Pi = Prob [u*i >-Xi] = F (Xi) (2)
where, F (Xi) is the cumulative density function (cdf).
The probit model is then derived by letting F (.) be the
cdf of a standard normal random variable. The
observed probit used to estimate the agro-forestry
products market participation function is specified as
yi= Xi + ei ; i = 1,., n (3)
where, yi is the dependent variable which takes a
value of 1 if the household participates in
agro-forestry products markets and value of 0 if the
household does not. is a vector of parameters to be
estimated. X is a vector of explanatory variables
hypothesized to influence the probability of
households participating in agro-forestry products
markets. e is the random disturbance term.
2.4.2 Model Specification
The Probit model was estimated as,
yi = Xi + ei ; i = 1,., n (4)
where,
 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
yi = Probability of market participation (1 = if the
farmer participated, 0 otherwise)
X is a vector of explanatory variables that include
X 1 = Age (Years)
X 2 = Sex of the farmer (1 = Male, 0 = Female)
X 3 = Education level (Years of formal schooling)
X 4 = Household size (Number)
X 5 = Farm size (Hectares)
X 6 = Market distance from farm (Km)
X7 = Group membership (1 = Member, 0 = Otherwise)
X 8 = Extension visits (Number of visits)
X 9 = Access to credit (1= Yes, 0 = Otherwise)
= Error term
= A vector of parameter estimates to be computed
2.4.3 Data Analysis
Primary data were entered in SPSS (Statistical
Package for Social Scientists) and analyzed using
Stata. Descriptive statistics in the form of
percentages, means and standard deviations were
generated to identify socio-economic characteristics
of farmers and types of agro-forestry products sold.
Comparison of socio-economic characteristics was
made using chi-square and t-test statistics for
categorical and continuous variables, respectively.
To analyze the factors influencing farmers
participation in agro-forestry markets, the Probit
model was used.
3. Results and Discussion
3.1 Socio-economic Characteristics of Market
Participants and Non-participants
Table 1 presents descriptive statistics of continuous
variables of agro-forestry farmers participating and
not participating in marketing their produce in Busia
District out of 153 farmers surveyed, only 50 farmers
participated in the marketing of agro-forestry products.
Agro-forestry farmers who did not participate in
marketing were on average older (43 years) than those
who participated and the difference was significant at
5% level of significance ( = 0.05). The mean
household size of market participants was
333
comparatively larger (with 8 members) as compared
to non-participants (with 6 members). The significant
relationship between household size and market
participation implies that farmers with a higher
number of household members were more likely to
participate in agro-forestry markets compared to those
with fewer members.
The average number of household members
actively involved in farming was higher (4 members)
for farmers who participated in marketing. This means
that households with a big number of members
actively involved in farming were more likely to
participate in agro-forestry markets because of
abundant labor supply provided by such members. On
the other hand, households with fewer numbers of
members involved in farming are less likely to
participate in agro-forestry markets because of labor
shortages, leading to low output in absolute terms.
Market participants held comparatively larger farms (3
ha) compared to non-participants (2 ha). In addition,
the education level of household heads in terms of
years of formal schooling was higher for market
participants (6 years) compared to 3 years for
non-participants. The significant relationship between
the level of education and market participation implies
that educated farmers participate more than the less or
non-educated ones. This is because education
enhances the ability to derive, decode and evaluate
useful information [28].
Results indicate that general farming experience in
terms of years was not significantly different across
the two groups of farmers. On the other hand,
experience in agro-forestry practices was significantly
different. Farmers who participated in agro-forestry
markets had more experience (of 4 years) in practicing
agro-forestry compared to their counterparts. Results
further indicated that the market distance between
market participants and non-participants was not
significant. On average, the two groups of farmers had
an average market distance of 8 km. In terms of
extension services, farmers who participated in
334 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics

Table 1 Descriptive statistics of continuous variables of agro-forestry farmers participating and not participating in marketing.
Market participants Non-participants
Variable (mean) t-value -value
(n = 50) (n = 103)
Age (years) 36 (11) 43 (12) 3.339 0.001
Household size (number) 8 (4) 6 (3) -2.494 0.014
Household members involved in farming (number) 4 (3) 3 (2) -2.979 0.003
Household members not involved in farming (number) 4 (3) 4 (2 ) 0.164 0.870
Education level of household head (years) 6 (4) 3 (4) -3.458 0.000
Distance to trading centre (km) 8 (6) 8 (6) 0.218 0.828
Farm size (ha) 3 (3) 2 (2) -3.243 0.002
General farming experience (years) 12 (10) 11 (10) -0.525 0.600
Agro-forestry experience (years) 4 (3) 2 (3) -3.546 0.000
Extension visits (number) 2 (2) 1 (1) -4.951 0.000

agro-forestry markets received more extension visits Table 2 Descriptive statistics of categorical variables of
agroforestry farmers participating and not participating in
(2) per month on average as compared to those who
marketing.
did not participate (1). This indicates that the more Non-part
Market
visits by the extension agents to the farmers, the more Variable (%) participants
icipants 2
-value -value
(n =
aware they become of the better agro-forestry (n = 50)
103)
marketing practices, hence participation in marketing. Sex
Table 2 presents descriptive statistics of categorical Male 78.0 83.5 0.680 0.410
Female 22.0 16.5
variables of agro-forestry farmers participating and
Major occupation
not participating in the marketing of agro-forestry Farming 96.0 94.2 0.226 0.634
products in Busia District. The results indicate that Non-farming 4.0 5.8
there were more male than female farmers Group membership
participating (78%) and not participating (83.5%) in Yes 70.0 40.8 11.499 0.001
No 30.0 59.2
marketing of agro-forestry products in Busia District.
Access to credit
This is because men tend to have better access to land, Yes 14.0 12.6 0.056 0.812
labor and other resources than women [29]. Farming No 86.0 87.4
was the major occupation for both farmers who Access to extension
Yes 52.0 24.3 11.646 0.001
participated and those who did not participate in
No 48.0 75.7
agro-forestry markets. A large percentage of farmers
(70%) who participated in agro-forestry markets agro-forestry markets between farmers having access to
belonged to farmer groups compared to their extension or not. Contacts with extension agents expose
counterparts. There is a significant difference (P = farmers to information, which stimulates market
0.001) in participation in agro-forestry markets between participation. Results further indicate low percentages
farmers belonging or not belonging to farmer groups. for both categories of farmers in terms of access to
Farmer groups are centers where market information credit.
can be accessed by farmers, thus enhancing
3.2 Types of Agro-forestry Products Marketed by
participation in markets. Similarly, in terms of access to
Farmers
extension services, farmers who participated in
agro-forestry markets had more access to extension Agro-forestry farming systems produce a wide
services compared to non-market participants. There is variety of crop, livestock and tree products. Table 3
a significant difference (P = 0.001) in participation in shows that the most commonly sold crop products by
 Linking Markets to Smallholder Agro-forestry Farmers as a 335
Strategy for Poverty Alleviation in the Tropics

Table 3 Crop products. Table 4 Trees and their products.

Crop output sold Frequency* (n = 49) Percentage (%) Frequency* Percentage
Trees and tree products
Maize 14 28.6 (n = 84) (%)
Cassava 11 22.4 Firewood 16 19.0
Coffee 10 20.4 Ovacado (Persea americana) 16 19.0
Beans 4 8.2 Jackfruit (Artocapus heterophylus) 14 16.7
Millet 3 6.1 Poles 10 11.9
Bananas 2 4.1 Musambya (Markharmia lutea) 8 9.5
Groundnuts 2 4.1
Musizi (Maesopsis eminii) 3 3.6
Sweet potatoes 2 4.1
Cotton 1 2.0 Kalitunsi (Eucalyptus species) 3 3.6
Omusolya (Makhamia spp) 3 3.6
*Multiple responses.
Mangoes (Mangifera indica) 1 1.2
agro-forestry farmers in Busia District are Maize *Multiple responses.
(28.6%), Cassava (22.4%) and Coffee (20.4%).
Table 5 Livestock and livestock products.
Apart from crops sold, agro-forestry farmers also Livestock and Frequency*
Percentage (%)
market a wide variety of trees and tree products. Table livestock products (n = 32)
4 indicates that farmers mostly sell firewood (19%) Goats 11 34.4
Birds 11 34.4
which is got from various agro-forestry tree species,
Pigs 5 15.6
including Eucalyptus species, Markharmia lutea, Cattle 3 9.4
Artocapus heterophylus, Milicia excelsa, Calliandra Sheep 1 3.1
calothyrsus and Sesbania sesban. This means that Milk 1 3.1

firewood is a major source of fuel for households in *Multiple responses.

Busia District. This finding is in line with the Busia Table 6 Probit model estimates of the determinants of
District State of Environment Report [25]. It is from market participation.
the same tree species that farmers obtain poles (11.9%) Variable dy/dx t-value P-value
Sex -0.056 -0.47 0.641
for sale. Fruits from tree species such as Persea
Age -0.019 -4.28 0.000***
americana (19%), Artocapus heterophylus (16.7%) Household size 0.026 1.85 0.065*
and Mangifera indicia (1.2%) are also sold by farmers. Formal education 0.021 2.09 0.037**
Makharmia lutea (9.5%), Maesopsis eminii (3.6%), Market distance -0.000 -0.01 0.989
Eucalyptus species (3.6%) and Makharmia spp (3.6%) Farm size 0.046 2.19 0.029**
Group membership 0.126 1.41 0.159
are sold as seedlings to other farmers.
Access to credit 0.236 1.71 0.088*
Agro-forestry systems in Busia District are Extension visits 0.103 3.21 0.001***
integrated with livestock, particularly goats, birds, *, **, *** Significant at 10%, 5% and 1%, respectively.
cattle and sheep. Table 5 indicates that goats and birds
(34.4%) are the most commonly sold livestock by of the nine variables included in the model, six were
agro-forestry farmers. These were followed by pigs significant in influencing farmers participation in
(15.6%), cattle (9.4%) and sheep (3.1%). Results agro-forestry markets. These were age, household size,
further indicate that milk (3.1%) was the only education level, farm size, access to credit and number
livestock product sold by farmers. of extension visits. Sex of the farmer, market distance
and group membership did not significantly influence
3.3 Factors Influencing Farmers Participation in
participation of farmers in agro-forestry markets at 5%
Agro-forestry Markets level of significance ( = 0.05). The marginal effects
The results of the Probit model of the determinants are indicated as dy/dx. These imply, for instance, that
of market participation are presented in Table 6. Out households who have had access to credit have a
336 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
23.60% higher chance of participating in agro-forestry
products markets than those who have no access to
formal credit. An increase in age by one year of the
household head reduces the probability of
participating in agro-forestry products markets by
1.90%. An increase in the household size by one
person increases the probability of participating in
agro-forestry product markets by 2.60%.
Age was significant at 1% and the sign is consistent
with the expectation that it influences market
participation negatively. Results indicate that an
increase in farmers age by one year would reduce the
probability of participating in agro-forestry markets by
1.9%, given that all other factors are held constant.
This implies that as a farmer grows older,
agro-forestry becomes a way of life rather than a
business. These results tally with those of Randela et
al. [30] who reported that younger farmers were
progressive, more receptive to new ideas and
understood the benefits of agricultural
commercialization compared to older farmers. Young
farmers have been recognized globally as savers of
environmental resources [31-34].
Household size was significant at 10%, with a
positive coefficient, implying that farmers with large
numbers of people in the household were more likely
to participate in agro-forestry markets. Results
indicate that as the number of active people in the
household increases by one person, there will be an
increase of 2.6% in the probability of participating in
agro-forestry markets. This could be due to the
availability of labor supply for farming activities
associated with large family sizes. In the study area,
the average number of household size participating in
the marketing of agroforestry products was
comparatively higher than that of non-market
participants (Table 1). Household size has been
reported to encourage market participation by other
scholars as well [35-37].
The positive and significant coefficient on the
education level of the farmer implies that educated
farmers were more likely to participate in the
marketing of agro-forestry products markets because
education improves farmers ability to understand and
interpret information. When all other factors are held
constant, an increase in the formal education of the
farmer by one year increases their probability of
participating in agro-forestry products markets by
2.1%, holding all other factors constant. Similarly,
Otieno et al. [38] reported that the intensity of market
participation significantly increased by the household
heads education level and access to formal market
information channels.
Farm size is a characteristic of the farm which has
continued to receive attention from agricultural
researchers [39]. Farm size was hypothesized to
influence market participation positively as the larger
the farm, the higher the probability of participating in
agro-forestry markets. Results indicate that an increase
in farm size by 1 hectare increases the probability of
participating in agro-forestry products markets by 4.6%.
Boughton et al. [40] and Randela et al. [30] also
reported positive and significant relationships between
farm size and market participation.
Results indicate that access to credit has a positive
and significant influence on the probability of
participating in agro-forestry products markets.
Households who have had access to credit have a
23.60% higher chance of participating in agro-forestry
product markets than those who have no access to
formal credit. A positive and significant effect of
access to credit was also reported by Randela et al.
[30]. There is a positive and significant relationship
between the degree of extension contact and
participation in agro-forestry products markets at 1%
level of significance. When all other factors are held
constant, an additional extension visit increases the
probability of market participation by 10.3%.
Farmers interaction with extension officers improves
their access to information and marketing skills. Thus,
the more extension visits to the farmers, the more
likely they would sell their produce [32-36, 41].
 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
4. Conclusions
This study revealed that only a few agro-forestry
farmers participate in the marketing of their
agro-forestry products. A combination of crop, livestock
and tree products is sold by farmers. Main crop
produces sold are Maize, Cassava and Coffee. Most
rural agro-forestry farmers in the district sell firewood,
in addition to fruits from Persea americana, Artocapus
heterophylus and Mangifera indicia. Markharmia lutea,
Maesopsis eminii, Eucalyptus species and Makhamia
spp. are sold as trees to other farmers. In terms of
livestock, farmers mostly sell live goats and birds.
Socio-economic factors that affect farmers participation
in agro-forestry markets include age, household size,
education level, farm size, access to credit and number
of extension visits. Policies aimed at improving the
quality and coverage of these variables are likely to
increase participation of farmers in agro-forestry
markets so as to improve their livelihoods.
Acknowledgments
The authors sincerely acknowledge the generous
financial support received from the National
Agricultural Research Organization (NARO) of the
government of Uganda that led to a successful
implementation of field activities. Authors thank the
farmers in Busia District of Uganda for actively
participating in this study. Authors also appreciate the
support of the College of Agricultural and
Environmental Sciences, Makerere University, more
especially the Principal Professor Sam Kyamanywa.
Authors are grateful to Professor Victor Ochwoh,
Makerere University, for the encouragement to carry
out fieldwork for this investigation and Mr. Albert
Muhumuza, the IT administrator at the College, for
his tireless efforts assisting us, and Special thanks to
Ms Ndibaisa Peruth Akiiki for the moral support.
References
[1] ICRAF (International Council for Research in
Agroforestry), World Agroforestry Centre, Southeast
Asia, 2006.
337
[2] P.K.R. Nair, An Introduction to Agroforestry, Dordrecht,
Netherlands, Kluwer Academic Publishers, 1993.
[3] J.R. Brandle, B.B. Johnson, T. Akeson, Field windbreaks:
Are they economical?, Journal of Production Agriculture
5 (1992) 393-398.
[4] T.R. Clason, Economic implications of silvipastures on
southern pine plantations, Agroforestry Systems 29 (1995)
227-238.
[5] W.B. Kurtz, Economics and Policy of Agroforestry,
2000.
[6] D.E. Rocheleau, Confronting Complexity, Dealing with
Difference: Social Context, Content and Practice in
Agroforestry, 1999.
[7] J.P. Lassoie, L.E. Buck, Development of Agroforestry as
an Integrated Land Use Management Strategy, 2000.
[8] B. Carson, The land, the farmer and the future: A soil
fertility management strategy for Nepal. ICIMOD
(International Centre for Integrated Mountain
Development) occasional paper No. 21, International
Center for Integrated Academic Publishers, 1992.
[9] P.A. Sanchez, Science in agro-forestry, Agroforestry
Systems 30 (1995) 5-55.
[10] B.T. Kang, F.K. Akinnifesi, Agroforestry as alternative
land-use production systems for the tropics, Natural
Resources Forum 24 (2000) 137-151.
[11] R.P. Neupane, G.B. Thapa, Impact of agroforestry
intervention on farm income under the subsistence
farming system of the middle hills, Nepal, Agroforestry
Systems 53 (1) (2001) 31-37.
[12] ICRAF (International Council for Research in
Agroforestry), Agroforestry for poverty reduction,
realizing the potential strategic plan for ICRAF East and
Central Africa Programme 2000-2010, 2003.
[13] PMA (Plan for Modernization of Agriculture), The
Quarterly Bulletin, Vol. 4, 2007.
[14] IFAD (International Fund for Agricultural Development),
Promoting market access for the rural poor in order to
achieve the millennium development goals, Discussion
Paper for the Twenty-Fifth Anniversary Session of
IFADs Governing Council, 2003.
[15] PORIA (Policy-Oriented Research Impact Assessment),
Case study on the International Food Policy Research
Institute (IFPRI) and the Mexican PROGRESA
Anti-poverty and Human Resource Investment
Conditional Cash Transfer Program, 2007.
[16] B.O. Muok, B. Owour, I. Dawson, J. Were, The Potential
of Indigenous Fruit Trees: Results of a Survey in Kitui
District, Kenya, Agroforestry Today, 1999.
[17] V. Levasseur, M. Djimde, A. Olivier, Live fences in
Segou, Mali: An evaluation by their early users,
Agroforestry Systems 60 (2004) 131-136.
[18] D. Russell, S. Franzel, Trees of prosperity: Agroforestry,
338 Linking Markets to Smallholder Agro-forestry Farmers as a
Strategy for Poverty Alleviation in the Tropics
markets and the African smallholder farmer, Agroforestry
Systems 61 (2004) 345-355.
[19] M.D. Byabashaija, J.F.O. Esegu, J.M. Kidiya, M. Basoga,
R.K. Ondia, Traditional uses of indigenous tree species,
Uganda Journal of Agricultural Sciences 9 (2004)
367-371.
[20] P. Kiwuso, G. Maiteki, J. Okorio, Indigenous methods of
controlling termites in agroforestry systems in Uganda,
Uganda Journal of Agricultural Sciences 9 (2004)
893-895.
[21] J.B.L. Okullo, J. Obua, G. Okello, Use of Indigenous
knowledge in predicting fruit production of Shea butter
tree in agroforestry parklands of North-Eastern Uganda,
Uganda Journal of Agricultural Science 9 (2004)
360-366.
[22] J. Okorio, J. Sekatuba, H. Agaba, C. Wandui, S. Musana,
G. Ongodia, et al., Potential of agroforestry in improving
livelihoods in Eastern and Mid-northern Uganda, Uganda
Journal of Agricultural Sciences 9 (2004) 901-907.
[23] B.M. Katumba, J.M. Boffa, G. Abigaba, J. Okorio,
Domestication of medicinal tree species in the Victoria
lakeshore region, Uganda Journal of Agricultural
Sciences 9 (2004) 84-88.
[24] UBOS (Uganda Bureau of Statistics), The 2002 Uganda
Population and Housing Census, Main Report, 2002.
[25] Busia District State of Environment Report, Published by
National Environment Management Authority, 2004.
[26] D.N. Gujarati, Basic Econometrics, 4th ed., McGraw-Hill,
Inc, 2004.
[27] W.H. Greene, Econometric Analysis, 4th ed., New
Jersey, Prentice Hall International Inc., USA, 2000.
[28] R.N. Uaiene, C. Arndt, W.A. Masters, Determinants of
agricultural technology adoption in Mozambique,
Discussion paper No. 67E, National Directorate of
Studies and Policy Analysis, Ministry of Planning and
Development, Republic of Mozambique, 2009.
[29] M.L. Morris, C.R. Doss, How does gender affect the
adoption of agricultural innovations?, The case of
improved maize technology in Ghana, International
Maize and Wheat Improvement Center (CIMMYT),
1999.
[30] R. Randela, Z.G. Alemu, J.A. Groenewald, Factors
enhancing market participation by small-scale cotton
farmers, Agrekon 47 (4) (2008) 451-469.
[31] P. Mayfroit, E.F. Lambin, Global forest transition:
prospects for an end to deforestation, Annual Review of
Environment and Resources, 2012. (in press)
[32] E. Kato, C. Ringler, M. Yesuf, E. Bryan, Soil and water
conservation technologies: A buffer against production
risk in the face of climate change?, Insights from the Nile
Basin in Ethiopia, Agricultural Economics 42 (5) (2011)
593-604.
[33] A. Flohre, M. Rudnick, G. Traser, T. Tscharntka, T.
Eggers, Does soil biota benefit from organic farming in
complex vs. simple landscapes?, Agriculture, Ecosystems
and Environment 141 (1-2) (2011) 210-214.
[34] B.E. Swanson, R. Rajalahti, Streghthening agricultural
extension and advisory systems: Procedures for assessing,
transforming and evaluating extension systems, World
Bank, 2010.
[35] S. Pascucci, T. de-Magistris, Effects of changing regional
agricultutral Knowledge and innovation system on Italian
farmers strategies, Agricultural Systems 104 (9) (2011)
746-754.
[36] M.T. Makhura, Overcoming transaction costs barriers to
market participation of smallholder farmers in the
northern province of South Africa, Ph.D. Thesis, Faculty
of Natural and Agricultural Sciences University of
Pretoria, 2001. (unpublished)
[37] M.L. Lapar, G. Holloway, S. Ehui, Policy options
promoting market participation among smallholder
livestock producers: A case study from the Philippines,
Food Policy 28 (2003) 187-211.
[38] D.J. Otieno, J. Omiti, T. Nyanamba, E. McCullough,
Market participation by vegetable farmers in Kenya: A
comparison of rural and peri-urban areas, African Journal
of Agricultural Research 4 (5) (2009) 451-460.
[39] CIMMYT, Modeling agricultural technology adoption
using the software STATA, Presented at a Training
Course Organized by CIMMYT-ALP for its NARS
Partners in Southern Africa on: Econometric Application
to Modeling the Adoption of Agricultural Technologies,
Harare, Zimbabwe, February 21-25, 2005.
[40] D. Boughton, D. Mather, C. Barrett, B. Benfica, D.
Abdula, Market participation by rural households in a
low-income country: An asset-based approach applied to
Mozambique, Faith and Economics 50 (2007) 64-101.
[41] L. Coetzee, B.D. Montshwe, A. Jooste, The marketing of
livestock on communal lands in the Eastern Cape
Province: Constraints, challenges and implications for the
extension services, South African Journal of Agricultural
Extension 34 (1) (2004) 81-103.
Journal of Agricultural Science and Technology B 2 (2012) 339-348
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Acacia mangium and Eucalyptus urophylla in Northern

Vietnam: The Relationship between Silvicultural Activity
and Quality of Stand

Chaw Chaw Sein, Hoang Lien Son and Ralph Mitlhner

Tropical Silviculture and Forest Ecology, Burckhardt Institute, Georg-August-Universitt Gttingen, Bsgenweg 1, Gttingen 37077,
Germany

Received: October 10, 2011 / Published: March 20, 2012.

Abstract: This paper sets out to determine the link between socio-economic attributes of smallholders, silvicultural activity and the
stand quality of Acacia mangium and Eucalyptus urophylla plantations in two unrelated communes in Phu Tho province, Vietnam.
The necessary data was collected by interviewing the smallholders and conducting inventory of the stands. The data was analyzed
with descriptive statistics, Mann-Whitney tests and Spearman correlations. More silvicultural practices were conducted by A.
mangium planters with larger areas. Most of the smallholders practiced the suggested silvicultural activities. Productivity was
affected by pruning in medium quality sites. This research recommends that smallholders were supported with incentives for the
plantation establishment, education for smallholders in silvicultural techniques, organization and development activities for the wider
market.

Key words: Acacia mangium, Eucalyptus urophylla, Northern Vietnam, silvicultural activity, quality of stand.

1. Introduction Carle [2] have recorded that the condition and yield of
woodlots in the tropics are seldom lower than their
The ability of plantations to produce goods and
possible amount due to an inappropriate species to
services makes them vital in reducing human pressure
sites, an insufficient seedling origin or even
on the natural environment. Currently, forest
unsatisfactory silvicultural management. Gunasena
plantations occupy only 5% of global forest cover
and Roshetko [3] and Byron [4] noted that either
(187 million hectares), yet they contribute 35% of
technical knowledge or training is occasionally
global round wood supply [1]. In many countries,
qualifying smallholders silvicultural activity.
forest plantations are the basis for world-scale forest
However, Byron [4] and Walters et al. [5] found that
products industries. Over the last two decades,
the attainment of tree planting or management is not
Vietnam has embarked on reforming the forest sector
only because of technical factors. Ravindran and
and has made efforts to shift from state controlled to
Thomas [6], Summers et al. [7] observed through their
more small-scale plantation forestry to reduce poverty
work in the tropics and subtropics that in addition to
and to meet a growing demand for industrial material.
the smallholders competence in knowing what is
Therefore, assessment of wood quality from present
required of them, socio-economic and motivational
smallholder plantation regions is critical. Varmola and
factors also have a distinct affect on their capacity for
tree planting.
Corresponding author: Chaw Chaw Sein, Ph.D., research
fields: silvicultural activity, stand quality, relationship, Acacia An example of the influence of socio-economic
mangium, Eucalyptus urophylla. E-mail: factors is the giving of incentives. Enters et al. [8]
chaw.chaw4@gmail.com.
340 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand
recommended that encouraging tree planting and
giving incentives persuade smallholders to indeed
plant the trees. But Haltia and Keipi [9] highlighted
that plantation management primary to great quality
and quantity of wood is uncertain when incentives are
the only motivation for smallholders to plant trees.
On the other hand, Bebbington [10] claimed that
participation in social organizations is undeniably
beneficial in encouraging farmers to learn better
farming practices. Moreover, Scherr [11] recommended
that if farmers are collectively organized, they can more
easily establish and manage their plantations and earn
much benefit from their great quality of wood. Walters
et al. [5] argued that comprehensive methods should be
used to learn the reasons for application of specific
practices by farmers. Yet the biography of farmers,
villages and precise procedures influence farmers
adjustment of practices [12].
The following study was deal on smallholders
silvicultural activities and practices in their plantations
in two communes in Phu Tho Province, Vietnam. The
purpose was to examine: (1) smallholders silvicultural
activity was affected by which socio-economic factors;
(2) which silvicultural practices presently applied by
smallholders; and (3) the influence of current
silvicultural practices to smallholder plantations. The
selected study sites deal with the research questions.
Byron [4] highlighted that smallholders tree plantings
can only be effective if clear land rights and
site-matching species, a practicable technology, crop
protection ability, and a wider market are in
appropriate. Blyth and Son [13] indicated that good
decisions are well-informed with reliable intelligence
on current and future market needs, product prices and
delivery costs, and credible technical advice on how to
meet market needs. The efficiency and effectiveness of
these supporting systems and services are critical to
good decision making.
To plan applicable tree planting schemes and
connected policies, perception of the socio-economic
and incentive parameters of smallholders and their
ideas for the quality of stand is valuable. Scherr [11]
argued that tree planting can potentially get better
income earnings if the farmers can produce wood
quality in accordance to the market demand.
The two species studied here were Acacia mangium
and Eucalyptus urophylla. Both species are exotic,
fast-growing tree species. A. mangium species are
incredibly multifunctional and can be used for
furniture and cabinet making, light to heavy
construction, mouldings, poles, posts, panelling,
mining timbers, boat building, carts, joinery, turnery,
tool handles, agricultural implements, matches, splints,
particle boards, hard board and veneer as well as for
plywood, pulp and paper [14]. The wood E. urophylla
is primarily used for pulp production, but it is also
used as a fuel-wood, for charcoal production,
electrical transmission poles, long-lasting posts and
pilings, in light and heavy construction,
cabinet-making, carpentry and for plywood and
agglomerate boards [14]. The recommended
silvicultural practices for both species include land
preparation, replanting and weeding, fertilising and
protection. In a review of pruning research on acacia
plantations, Dung et al. [15] concluded that after three
years only height growth was significantly different
between the pruned and unpruned treatments. It is
possible that pruned branches in the lower shaded part
of the crown were not providing enough photosynthetic
energy. For E. urophylla, pruning is not necessary as it
is self-pruning [16]. Thinning is not necessary for
either species as the end product is for pulp and paper
production companies located in the neighbouring
provinces. In this research, A. mangium and E.
urophylla plantations were maximum 7 and 6 years
old.
2. Materials and Methods
2.1 Study Sites
Two study areas Tram Tham and Canh Dinh
communes were selected. They are located in the
province of Phu Tho, Northern Vietnam. The area is
 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand
3,519 square kilometers, lying between 2055 to
2143 N and 10448 to 10527 E. The important was suggested for A. mangium. Defect caused by
earnings of both communes are agriculture, tea and
pulp and paper production. The sites were selected as
smallholders are planting trees with energetic efforts
for pulp and paper production. Most of the
smallholders had received A. mangium and E.
urophylla seedlings from the collaborative companies.
Smallholders have to sell all productivity to the
company and their benefits depend on the company
investments with different amounts and kinds of
inputs, labour, technical guidelines, and sometimes
even land.
2.2 Research Methods
Questionnaires dealing with socio-economic and
silvicultural characteristics were built. Interviews were
carried out at smallholders houses by well trained
people assigned by the local Forest Research Institute
of Vietnam. Interviewees were questioned on: (1)
farm and household characteristics, such as number of
household members, age, (2) education, and (3)
occupation, land possession, earning from plantation,
assets from loans and participation in farmers groups
or other social organizations. They were also asked
about: (1) criteria for selection of species, the present
silvicultural practices conducted by smallholders; and
(2) incentive and attitudes on tree planting by
smallholders. By interviewing several smallholders,
questionnaires were tested and some were modified.
Data were collected from 14 smallholders for A.
mangium and 24 smallholders for E. urophylla stands,
respectively.
Using Spearman correlations, descriptive statistics,
Mann-Whitney tests and cross tabulations [17], data
were analyzed to learn the links between
socio-economic characteristics and silvicultural
activity applied by smallholders. The silvicultural
activity was decided in grade from one to seven for
each stand. For A. mangium stands, the four practices
applied by smallholders were land preparation,
341
refilling, weeding and fertilizing. In addition, pruning
natural or human disturbance to development of
stands was suggested to protect for both species even
though it was not included in the analysis.
The data relating to the required stands were
gathered by visiting the local commune office of Phu
Tho Province. Inventory was carried out in all the
stands of the selected smallholders and each stand was
classified using optical inspection. The stand was
subdivided into two parts by cutting the longest
diagonal into two by a perpendicular bisector. In each
of the two parts of the stand, one plot center was
determined randomly. When the stand was too big,
more than two plots were allocated. A fixed circular
plot was laid out to encompass sample trees on which
observations were done. A high number of trees per
hectare lead to a smaller plot size and a low density to
a larger one. A standard rule of thumb, achieving an
average of about 20 trees per plot was used. Often the
plots have been made larger to get more data. The
measurements, diameter, total tree height, stem quality
and overall tree condition were then taken and the
height of the five nearest trees to the sample points
were measured with an electronic hypsometer and
telescope stick. The remaining trees were estimated
using the height-diameter relationship equation:
Trees were classified into three classes: (1)
dominant; (2) co-dominant; (3) suppressed trees. The
link between plantation quality and the silvicultural
practices applied by the smallholders was examined
with descriptive statistics, hypothesizing and
statistically fitting a general linear model. By using
probable productivity as quality indicator, plot quality
affected by silvicultural practices was examined. The
study was conducted at the plot level due to various
tree density and site quality of plots within a stand of
same age and practices. As most of the farmers
conducted land preparation, fertilizing and weeding in
their plots, there was no variation in productivity
among plots due to these activities. Pruning was
342 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand

chosen as an indicator to examine any effect on A.

mangium plot quality. According to Son et al. [18],
pruning is an essential practice to avoid breakage of
stems by heavy rains. Due to various protection types
and requirements, it was not considered in this study.
The effect of replanting was also not considered as the
various mortality rates among the plots. A general
linear model (n = 104) was hypothesized to analyze
whether there was an effect on the potential
productivity by pruning. The null hypothesis showed
that productivity was not affected by pruning, site
quality, stand age, and tree density. In the model, site
quality and pruning treatment were considered as Fig. 1 Silvicultural practices applied by smallholders
fixed factors while plot age and tree density were the plantations.
covariates. Number of stands: Acaci mangium = 14, Eucalyptus urophylla
= 24.
3. Results Table 1 Source of silvicultural practices (learnt by
smallholders in study sites).
3.1 Silvicultural Practices and Activity Conducted by
A. mangium E. urophylla Total
Smallholders Source
F RF(%) F RF(%) F RF(%)
Other tree planters 3 21 6 25 9 24
According to the results of the questionnaires, Public institution 1 7 4 17 5 13
almost all of the smallholders undertook more than Company 1 7 3 13 4 11
two thirds of the silvicultural practices. In A. mangium Farmers
4 29 7 29 11 29
association
stands, the average grade was 2.0 (SD = 1.32) and Own experience 5 36 4 17 9 24
3.43 (SD = 2.17) in E. urophylla stands out of 7 Total 14 100 24 100 38 100
Number of Acacia mangium planting smallholders n = 14,
silvicultural activities for both (Fig. 1). The age of
Eucalyptus urophylla planting farmers n = 24. F = frequency.
plantation, selection of species and the use of the RF = relative frequency.
products influenced the silvicultural practices used.
Table 2 Specific silvicultural practices conducted by
Nearly a third of the smallholders who applied smallholders (%).
silvicultural practices relied on their own experience Management A. mangium E. urophylla
Total (%)
and half relied on information from their farmers activity (%) (%)
Land preparation 100 88 92
association and from the neighbouring tree planters Fertilizing 57 54 55
(Table 1). Most of the smallholders laid out the land, Refilling 64 63 63
fertilized it and set about protecting it heedless of the Weeding 71 71 71
Pruning 79 21 79
species planted (Table 2). It was observed that only Protection 79 92 79
55% of the smallholder used fertilizer, even though Thinning 0 0 0
this is a necessary practice for the improvement of Number of Acacia mangium planting smallholders n = 14,
Eucalyptus urophylla planting smallholders n = 24.
plantation. This was identified as a key dilemma with
tree planting caused by the relatively expensive replanting was done in approximately 64% of the A.
fertilizers especially in E. urophylla plantations. mangium and 63% of the E. urophylla stands.
Although the mortality rate was just over 25% of A. Compatively energetic effort of replanting exhibits the
mangium stands and only 21% of E. urophylla stands, farmers involvement in stand productivity.
 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The 343
Relationship between Silvicultural Activity and Quality of Stand

Most of the smallholders relied on their own Table 3 Smallholders and their plantation (qualitative
characteristics).
knowledge and experience when deciding how to
A. mangium E. urophylla Total
prune their stands. The average age range for the first Characteristic
n Yes % n Yes % n Yes %
pruning was from 1 to 2 years. Pruning was vital in 1 14 86 24 96 38 92
order to get rid of deficient branches caused by insects, 2 14 100 24 100 38 100
fire or wind. Thinning practices were not found in the 3 13 38 22 36 35 37
4 14 93 23 96 37 95
smallholder plantations. Thinning was not actually
5 14 50 24 75 38 66
required in the stands because these were fast growing
6 12 67 24 83 36 78
species, their rotation is too short and the end product 7 14 36 24 33 38 34
is for pulp and paper production. Almost all of the 8 14 79 24 79 38 79
farmers carried out protection activities and fire 9 14 57 24 63 38 81
10 14 57 24 71 38 66
prevention within the whole year. The rotation length
Number of Acacia mangium planting smallholders n = 14,
of A. mangium was observed ranging from 5 to 7 Eucalyptus urophylla planting smallholders n = 24.
years and for E. urophylla from 4 to 6 years. The Characteristics:
primary factor mentioned in determining rotation 1 = Decided themselves suitable species;
2 = Conducts silvicultural management themselves;
length was how long it took the trees to reach a
3 = Can measure stand yield;
merchantable size. Two thirds of the farmers indicated 4 = Has clear land tree plantation;
that they did not have enough knowledge to predict 5 = Received seedlings as incentives;
the volume (Table 3). Most of the smallholders did 6 = Received financial incentives;
7 = Obtained training;
not have a scheduled harvesting age due to unstable
8 = Described financial incentives for tree planting;
economic circumstances and the fluctuating market 9 = Has favourable or highly favourable attitude towards tree
price of the buyers. planting;
Over half of the smallholders practiced their 10 = participated in Farmers group.

management activities in concert with other tree
planters. The A. mangium planting smallholders
mainly cooperated in fire protection. More than half of
the farmers had knowledge to apply further activities
with neighboring tree planters.
3.2 Fundamental Parameters Affecting the Grade of
Silvicultural Activity of Smallholders
Potential links between the grade of silvicultural
activity and the socio-economic characteristics of
smallholders and stand were examined. In A. mangium
stands, it was observed that smallholders planting
more than one stand also applied more silvicultural
practices. In Table 4, the same correlation signs
(between the species) of the grade of silvicultural
activity and the socio-economic variables in 7 of the
10 cases were examined. A. mangium planting
smallholders with less household possessions and less
income from plantations conducted more silvicultural
practices, while E. urophylla planting smallholders
with more household possession and more earnings
from plantations but undertook less silvicultural
activities.
For A. mangium planting farmers, percentage of
planted area of total land area, total income from
plantation, years of education, other peoples
influence, distance from plantation to smallholders
house and household possession borrowed from the
bank were negatively correlated with smallholders
silvicultural activity. A significant difference was
noticed about the grade of silvicultural activity for A.
mangium planters contributing towards total farm area,
percentage of total land area for plantation and
household possessions borrowed from the
smallholders bank. The age of the smallholders,
number of household members, total plantation area,
344 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand

Table 4 Spearman correlations for grade of silvicultural plantation contrast to the E. urophylla planters.
management and socio-economic characteristic of
Comparison between the groups of A.mangium and
smallholders.
E. urophylla planting smallholders, they had
A. mangium planters E. urophylla planters
n Mean SD r n Mean SD r analogous characteristics relating to land ownership,
1 14 50.43 12.65 0.18 24 49.08 14.18 -0.08 incentives for planting, species selection, obtainable
2 14 4.21 1.05 0.22 24 4 0.78 0.22 incentives and communication with farmers group or
3 14 3.43 0.94 0.49* 24 3.3 0.96 0.45
another social organisation (Table 3). Therefore, the
4 14 1.79 0.7 0.21 24 2.17 0.76 0.13
5 14 1.74 1.15 0.03 24 2.46 5.2 0.06
difference in smallholders silvicultural activity
6 14 52.72 31.44 -0.58* 24 3.27 1.02 -0.52 probably has not resulted from these characteristics.
7 14 289.45 467.37 -0.42* 24 294.95 540.96 0.05 As the smallholders commonly adhered to the
8 14 1.43 0.73 -0.29 24 1.62 0.68 -0.49 characteristics described above, these results were not
9 14 300.21 368.71 -0.10 24 532.67 725.85 -0.1
statistically analyzed.
10 14 7.07 1.77 -0.06 24 6.54 1.61 -0.09
11 14 1.85 0.86 -0.32 24 2.21 0.78 -0.51 All smallholders interviewed were the legal
Number of Acacia mangium planting smallholders n = 14, proprietors of their tree stands and identified financial
Eucalyptus urophylla planting smallholders n = 24, * P < 0.05. incentives for tree planting, but the choice of species
1 = Age of smallholders (years);
was also subjected by the company supporting for the
2 = Number of household members;
3 = Total land area (ha); plantation establishment. Attitudes towards tree
4 = Number of income earnings household members; planting among A. mangium planting smallholders
5 = Total area planted by Acacia/Eucalyptus (ha); showed that almost 40% had indifferent or
6 = Percentage of planted area of total land area (%);
unfavourable attitudes towards tree planting. Likewise
7 = Household possessions from bank;
8 = Distance from plantation to farmers house (km); for E. urophylla, 60% were found to exhibit
9 = Total earnings from plantations (USD (2007)/year; favourable or highly favourable attitudes while 40%
10 = Education (years of study); were reported as being indifferent or unfavourable
11 = Other peoples influence.
towards tree planting. In interviews, nearly half of the
total earnings from the plantations, years of study, smallholders described uncertain markets or variable
number of income earning household members, other sources of incomes as disadvantages of tree planting.
peoples influence and distance between plantation A. mangium planting smallholders showing a
and the smallholders house were not significantly favourable or highly favourable attitude towards tree
correlated with the grade of silvicultural practices. planting more seldom conducted all four required
However a contrast of descriptive statistics among silvicultural practices than smallholders with
the two smallholders groups is illuminating, the indifferent or unfavourable attitudes (Table 5).
variation between the socio-economic characteristics In this study, over half of the smallholders were
of the A. mangium and E. urophylla smallholders was members of a farmers group; consequently,
not statistically examined. Table 4 shows that A. participation in the farmers group did not necessarily
mangium planters were on average older, and had clarify the selection in smallholders silvicultural
more household members, less income-earning activity (Table 3). A. mangium planting smallholders
household members, more total land area, a larger participated in a farmers group for a longer time was
percentage of plantation area, plantations located enthusiastic in conducting silvicultural practices.
nearer to their houses, lesser total earnings, lesser Additionally, farmers group helped neighboring
household possessions from the bank, more years of smallholders in case of applying silvicultural practices.
education and less influenced by other people for tree About half of the E. urophylla and A. mangium planters
 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The 345
Relationship between Silvicultural Activity and Quality of Stand

Table 5 Acacia mangium planting smallholders: Cross Table 6 Acacia mangium planting farmers: Cross tabulation
tabulation of number of silvicultural practices with of grade of silvicultural management with position of training
smallholers attitudes towards tree planting. obtained for tree planting.
Favourable or Indifferent or Received training for tree planting
Grade of Total
highly favourable unfavourabe Grade Yes No Total
management
F RF (%) F RF (%) F RF(%)
F RF(%) F RF (%) F RF (%)
One practice 1 13 0 0 1 7
One practice 0 0 1 11 1 7
Two practices 1 13 0 0 1 7
Two practices 0 0 1 11 1 7
Three practices 0 0 1 17 1 7
Three practices 0 0 1 11 1 7
Four practices 2 25 1 17 3 21
Four practices 1 20 2 22 3 21
Five practices 2 25 2 33 4 30
Five practices 2 40 2 22 4 29
Six practices 1 13 2 33 3 21
Six practices 1 20 2 22 3 21
Seven practices 1 13 0 0 1 7
Seven practices 1 20 0 0 1 7
Total 8 100 6 100 14 100
Number of Acacia mangium planting farmers n = 14,
Number of Acacia mangium planting smallholders n = 14,
Eucalyptus urophylla planting farmers n = 24.
Eucalyptus urophylla planting smallholders n = 24.
Table 7 Summarizing of inventory data for two selected
stated that they had learnt their silvicultural practices species.
from other tree planters and from the farmers group A. mangium E. urophylla
Variable
(Table 1). Smallholders could probably obtain Mean Min Max Mean Min Max
seedlings from companies or by collecting themselves. 1 1.43 0.5 2.5 0.68 0.5 1.63
2 3.99 1.32 7.0 3.13 1.32 5.57
Farmers who obtained training always applied four
3 9.7 2.1 17.5 6.0 1.6 11.9
or more of the suggested silvicultural practices, 4 10.7 2 21.4 8.7 1.9 16.4
whereas A. mangium smallholders obtaining training 5 60.7 0.8 197.4 40.6 0.5 209.0
conducted fewer practices (Table 6). Nevertheless, the 6 2,072 1,333 3,333 2,546 1,333 4,444
7 1,267 508 2,663 2,603 486 5,856
silvicultural activity within the groups of smallholders
8 24.50 6.50 61.90 20.82 4.28 49.18
obtaining training was merely the implementation of Number of plots measured: Acacia mangium n = 104,
one third of these practices. A. mangium planting Eucalyptus urophylla n = 110.
smallholders obtained training techniques for seedling 1 = Distance from farmers house to stand (km);
preparation and planting. 2 = Age of stands (years);
3 = Tree diameter 1.3 (cm);
3.3 Stand Quality 4 = Tree height (m);
5 = Stand volume (m3/ha);
In Table 7, the growth characteristic is summarized 6 = Planting density (trees/ha);
7 = Actual density (trees/ha);
as mean, minimum and maximum values of all the
8 = Mortality (%).
measured plots by species. For A. mangium stands
(104 plots), they achieved a minimal increase in quality (Table 8). For A. mangium plots, the mean
diameter of 2.1 cm and height of 2.0 m at 1 year of percentage of stems with high quality was
age and a maximum increase in diameter of 17.5 cm approximately 57%, and for E. urophylla approximately
and height of 21.4 m at 7 years of age. For E. 25%. The high quality stems occurred in plots with
urophylla stands (110 plots), they achieved a minimal higher proportions of dominant trees (P = 0.000).
increase in diameter of 1.6 cm and height of 1.9 m at 1 The high growth of stands was described by both A.
year of age and maximum increase in diameter of 11.9 mangium and E. urophylla planting smallholders. 42%
cm and height of 16.4 m at 6 years of age. of the observations were the same for comparing the
The A. mangium stands were observed to be of a field teams with the observations on stand quality,
high quality and the E. urophylla stands of medium 38% of the observations differed by one quality class
346 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand

Table 8 Stand qualities of two selected species. 140

Observed A. mangium E. urophylla 120

stand quality F RF (%) F RF (%)
Low 2 14 13 21 100

Productivity(m /ha)
Medium 4 29 5 54

3
80
High 8 57 6 25
Total 14 100 24 100 60

Number of plots measured: Acacia mangium n = 104,

40
Eucalyptus urophylla n = 110.
20

and 20% differed by two quality classes. Individual

smallholders used to check the quality of their stands
0
Pruned Unpruned

Plot
by comparing other smallholders belonging to many
stands with energetic effort. (a)
120

3.4 The Currently Conducted Silvicultural Practices

100
and Quality of Stand
Productivity(m /ha)
80
It was observed that the potential productivity of
3

standing wood was not affected by pruning or 60

un-pruning (P = 0.534), site quality (P = 0.370), plot 40

age (P = 0.311) or tree density (P = 0.635). In

higher-quality sites, higher potential productivity 20

was occurred among the pruned and the un-pruned 0

high low mediu
plots (Fig. 2a). The productivity of pruned stands Sitequality

was lower in high quality sites and higher in medium (b)

and lower quality sites in contrast to un-pruned
stands (Figs. 2a and 2b). Pruning was contributed to
the productivity of the medium quality sites.
Additionally, scheduled of pruning and growth
recovery time after pruning, the pruning method,
intensity of pruning or percentage of dominant or
co-dominant trees in the plot may cause variation
depending on plot age and planting densities. The
effect of scheduled time of pruning in different site
qualities could not be examined as the study did not
have sufficient information for analysis.
4. Discussion
For A. mangium planting smallholders,
smallholders belonging to the largest plantation area
applied the majority of silvicultural practices. The
practicality of allocating land and labor to silvicultural
management were also influenced by market accessibility.
Fig. 2 (a) Productivity of pruned and un-pruned plots for
three site qualities. (b) Productivity in high quality sites,
medium quality sites and low quality sites in pruned and
un-pruned plots
In E. urophylla planting sites, the market was timid;
smallholders with adequate resources had
intercropped with some cassava. In the case of A.
mangium the number of household members earning
income was positively correlated with their
silvicultural activity, meaning that more household
members were able to work on the farming activities.
In this study, several A. mangium planting
smallholders who participated in the farmers group for
a longer period were more energetic efforts in
silvicultural management. Additionally, it was found
that several smallholders learned their silvicultural
practices from other tree planters or from the farmers
group.
 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand
Mahapatra and Mitchell [19] observed that farmers
tree planting activities were affected by their attitudes
to tree planting. In this study, A. mangium planting
smallholders with a favourable or highly favourable
attitude to tree planting undertook more silvicultural
practices. Salam, et al. [20] argued that extension
activities by foresters could have a positive effect on
tree planting. E. urophylla planters who obtained
training often carried out more suggested silvicultural
activities while smallholders who did not obtain
training carried out fewer.
According to the study, almost half of the A.
stands and a quarter of the E. urophylla stands were
observed to be of high quality. Approximately similar
observation was found on stand quality accessed by
field teams and those of the smallholders. The volume
of potential productivity varied more due to different
rotation, density and site qualities rather than by
particular silvicultural activity applied by
smallholders.
5. Conclusion
This study highlights that the silvicultural activity,
socio-economic and market-related characteristics of
each site are very important to accomplish in tree
planting. Prior to implementing local socio-economic
and wood based market of the sites should be
carefully inquired. It is also essential to record that the
positive effects of a certain silvicultural management
technique at one particular site may have a negative
effect at another site. Most existing growers originally
planted Acacia and Eucalypt species to supply
Vietnams pulp and paper industry. The timber market
availability appears having a tough contact on the
level of silvicultural management activity in the area.
The prices for sawn timber are increasing under the
influence of strong domestic and international demand
for wood products. The Acacia supply chain is a
multi-channel, multi-product system-presenting grower
with many choices. However, with the support of
further research on scheduled time, frequency and
347
silvicultural management techniques can probably be
improved the quality of stands. According to interview
results, farmers believed that they might achieve better
outcomes by improving silvicultural management.
There is also one possibility to improve the silvicultural
management of the quality of stands by introducing
fertilizing schemes experiments. Another potentially
beneficial approach would be the provision of training
in A. mangium pruning and volume estimation
techniques. It is obvious that procedures to improve
quality of stands may only be reasonable if a solution
can initially be organized for A. mangium planting
smallholders to market their wood.
There are some issues influencing silvicultural
management that are common to many similar
situations in developing nations around the world.
Adequate policies for incentives to plant species and
accessible markets for smallholders are needed. Since
the smallholders plant trees mainly for financial
reasons, this is problematic without certainty of future
income. In conclusion, it is necessary to develop
mechanisms to facilitate farmers in organizing
themselves and thereby jointly improve their
silvicultural activities of the stands and accessibility of
wider markets.
Acknowledgments
Authors would like to express their sincere thanks
to the staffs of the Silviculture Division, Forest
Science Institute of Vietnam for the great help in data
collection. In addition, authors specially thank to
Professor Dr. Christoph Kleinn, Mr Sebastian Schnell
(M.Sc.), Dr Christoph Fischer, all from Institute of
Forest Inventory and Remote Sensing, University of
Gttingen and all who have provided invaluable
contributions to the development of this manuscript.
References
[1] ICFPA, Statements on Forest Plantations, Ottawa,
Canada, 2005.
[2] M.I. Varmola, J.B. Carle, The importance of hardwood
plantations in the tropics and sub-tropics, Int. For. Rev. 4
348 Acacia mangium and Eucalyptus urophylla in Northern Vietnam: The
Relationship between Silvicultural Activity and Quality of Stand
(2) (2002) 110-121.
[3] H.P.M. Gunasena, J.M. Roshetko, Tree Domestication in
Southeast Asia: Results of a Regional Study on
Institutional Capacity, International Center for Research
in Agroforestry (ICRAF), Bogor, 2000.
[4] Byron, Keys to smallholder forestry, Forests Trees
Livelihoods 11 (4) (2001) 279-294.
[5] B.B. Walters, C. Sabogal, L.K. Snook, E. Almeida,
Constraints and opportunities for better silvicultural
practice in tropical forestry: An interdisciplinary
approach, For. Ecol. Manag. 20981-2 (2005) 3-18.
[6] D.S. Ravindran, T.H. Thomas, Trees on farms, stores of
wealth and rural livelihoods-insights and evidence from
Karnataka, India, Int. For. Rev. 2 (3) (2000) 182-190.
[7] P.M. Summers, J.O. Browder, A.M. Pedlowski, Tropical
forest management and silvicultural practices by small
farmers in Brazilian Amazon: Recent farm-level evidence
from Rodonia, For. Ecol. Manag. 192 (2-3) (2004)
161-177.
[8] T. Enters, P.B. Durst, C. Brown, J. Carle, P. Mckenzie,
What Does It Take?, The Role of Incentives in Forest
Plantation Development in Asia and the Pacific,
Asia-Pacific Forestry Commission, Regional Office for
Asia and the Pacific, FAO, Bangkok, 2004.
[9] O. Haltia, K. Keipi, Financing forest investment: The
issue of incentives, in: K. Keipi (Ed.), Forest Resource
Policy in Latin America, Inter-American Development
Bank, Washington, DC, 1999, pp. 63-78.
[10] Bebblington, Social, capital and rural intensification:
Local organizations and islands of sustainability in the
rural Andes, Geogr. J. 163 (2) (1996) 189-197.
[11] S.J. Scherr, Building opportunities for small-farm
agroforestry to supply domestic wood markets in
developing countries, Agrofor. Syst. 61-62 (1) (2004)
357-370.
[12] B.B. Walters, A. Cadelina, A. Visitacion, E. Cardano,
Community history and rural development: Why some
farmers participate more readily than others, Agric. Syst.
59 (2) (1999) 193-214.
[13] M.J. Blyth, H.L. Son, Socio-economic factors influencing
smallholder production of acacia sawlogs in Viet. Nam.,
ACIAR Project, FST (025), 2007.
[14] PROSEA, Timber trees: Minor commercial timbers, in: I.
Soerianegara, R.H.M.J. Lemmens (Eds.), Plant Resources
of South-East Asia, Wageningen, The Netherlands, 1995.
[15] P.T. Dung, N.V. Ngoc, H.V. Phuc, N.T. Le, N.T. Nhun,
P.V. Tung, et al., Research on Techniques of Intensive
Establishment for Selected Clones of Acacia Hybrid in
the Old Alluvial Soil in Binh Phuoc Province for Paper
Material, FSIV, Hanoi. Project Report, 2005.
[16] M.R. Jacobs, Growth Habits of the Eucalypts, Forestry
and Timber Bureau, Canberra, 1955.
[17] E. Ranta, H. Rita, J. Kouki, Biometria: Tilastotiedett
Ekologeille, 5th ed., Yliopistopaino, Helsinki, 1994.
[18] N.H. Son, N.X. Quat, D.H. Nam, Planting Techniques of
Tree Species for Raw Materials, Statistics Publishing
House, 2006, p. 128.
[19] A.K. Mahapatra, C.P. Mitchell, Classifying tree planters
and non-tree planters in a subsistence farming system
using a discriminant analytical approach, Agrofor. Syst.
52 (1) (2001) 41-52.
[20] M.A. Salam, T. Noguchi, M. Koike, Understanding why
farmers plant trees in the homestead agroforestry in
Bangladesh, Agrofor. Syst. 50 (1) (2000) 77-93.
Journal of Agricultural Science and Technology B 2 (2012) 349-361
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Assessment of Grain Damage and Weight Loss on Farm

Stored Maize in Highlands Areas of Bungoma District,
Kenya

Kimondo Mutambuki and Christopher Mugo Ngatia

Kenya Agricultural Research Institute, National Agricultural Research Centre, P.O. Box 14733-00800, Nairobi, Kenya

Received: September 13, 2011 / Published: March 20, 2012.

Abstract: A simulation trial to assess losses in crib and in-house maize storage practices was conducted. Three cribs were
constructed and 180 kg maize in two bags treated with 1.6% pirimiphos methyl and 0.3% permethrin stored in one. In the other, a
similar quantity of untreated maize in two bags was stored while in the third crib, 4 bags of cob maize were placed. Another four
bags, two treated and two untreated were placed on dunnage separated by a 2-meter path in the living house. Samples were analysed
at 4-week interval for 24 weeks and pest damage was used to calculate percent weight loss. Grain from selected farmers was also
analysed for comparison. Crib trial had 53% and 20% pest damage in untreated and treated maize compared with 56% and 16% for
in-house storage after six months. Cumulative weight loss averaged 20.6% and 9.7% for untreated and treated maize. The figures are
markedly higher than 5% criterion for effective pest control measures. Weight loss on farmer stocks compared well with untreated
maize and grain damage varied from 5.4% to 27%, translating to 0.5% and 16.8% cumulative loss. These results suggest ineffective
pest control measures by Bungoma farmers. The results contrast with Kitui ones in an earlier study where Prostephanus truncatus is
endemic and contributed to 10% cumulative loss above the Bungoma level. Efforts should be made to reduce grain damage to below
5% to realise food security at farm level.

Key words: Bungoma, consumption, insect pests, losses, maize (Zea mays L.).

1. Introduction important insect pests in stored maize grain being

Prostephanus truncatus Horn, Sitophilus zeamais
Maize (Zea mays L.) is Kenyas most important
Motschulskay and Sitotroga cerealella Oliver [9, 10].
staple crop with annual production fluctuating between
The entry and establishment of the larger grain
2.4-3.2 million tons in 2009 and 2010 respectively [1].
borer (LGB), Prostephanus truncatus Horn in Kenya
About 80% of this maize is stored on the farm by small
for almost 20 years has introduced a new dimension in
scale farmers allowing only 20% to be stored in the
the levels of storage losses for maize.
central storage system. Globally, post harvest grain
The real losses caused by P. truncatus to
losses are caused by insect damage and other
unprotected maize during one storage season have
bio-agents ranging from 10%-40% [2, 3]. In many
been estimated to be between 10% and slightly over
tropical countries, storage of grain for long periods is
30% by weight, significantly more than 2-3 times of
adversely affected by its susceptibility to weevil
what is normally lost as a result of attack by
damage [4-7]. In Kenya, the main causes of losses in
indigenous insect pests [11, 12]. Golob et al. [13]
durable crops are insects and rodents [8] with the most
recorded 34% loss in Tanzanian maize stored on cob
with husk for 3-6 months while in Kenya, losses due
Corresponding author: Kimondon Mutambuki, M.Sc.,
MBA, research fields: postharvest managements in durable to P. truncatus was 35% [14].
grains, mycotoxins, vegetables and also agribusiness. E-mail:
mutambukikimo@yahoo.com.
A survey was carried out in 2001 in Bungoma, an
350 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
area endowed with high maize production potential, to
assess post-harvest losses at farm level. As an output,
groups of farmers prioritised maize storage as a major
constraint that needed to be addressed, which were
collaborated by individual farmer interviews. The
constraints then formed the objective of the on farm
simulation trial set up to assess weight losses on
stored maize due to pest infestation with and without
intervention measures. The results were to be
compared with losses incurred in a few selected
individual farm stores and farmers coping strategies.
The survey had also shown a storage pattern where
maize was temporarily stored in cobs in outside cribs
and later transferred into a living house store after
shelling and treating with chemical dusts. Hence, the
influence of the crib and house environments on the
impact of storage pests on maize was also assessed.
2. Materials and Methods
2.1 Geographical Description of the Site
Bungoma is about 410 km to the North West of
Nairobi in Western Province. It is situated on the
slopes and foothills of Mount Elgon. The altitude
varies from 1,200 m in the lower midland zones to
3,500 m in the tropical Alpine zones. The average
annual rainfall ranges from 1,000-1,800 mm with two
rain seasons in a year. There is only one crop of maize
in a year. Bungoma is divided into the following
agro-ecological zones:
Tropical Alpine-TA, TA1, TA2
Upper Highlands-UH, UH0, UH1
Lower Highlands-LH1, LH2
Upper MidlandsUM2 and UM3
Lower Midlands-LM1, LM2, LM3
The annual mean temperature in LM3 ranges from
21 C-22 C while in northern parts TA1&2, it varies
from 5 C-10 C [15]. The site of the trial was in UM2
but the farmers were spread out in UM2 and UM3.
2.2 Farmer Sampling
In the district, a different locality not covered
during the survey was selected in order to ensure that
any additional opinion or information was
independent. Six farmers with a history of high maize
production who were likely to store for at least six
months were randomly recruited from each district. At
the homestead, the quantity of the maize found in the
store was recorded in terms of 90 kg bag equivalents.
Additional information on the month of harvest, maize
treatment with pesticides, frequency of maize cooking,
and family members who ate from the home was
recorded.
Before monthly samples were taken, the bag from
which the portion for the family use was drawn was
first weighed using an Avery balance of 1.0 kg
accuracy (W&T Avery Ltd., UK). The weight
difference during the subsequent visit was deemed to
be the quantity used by the family for that month. To
determine monthly consumption pattern by each
family, a quantity of maize that was cooked in a day
was weighed in local plastic containers. The amount
consumed was calculated on weekly basis and later
converted to monthly figures with adjustments for the
days when maize was not cooked. For each sample
taken during sampling, 2 kg of treated maize were
given as a replacement.
2.3 Simulation of Farmer Storage Practice
On the site, two traditional cribs which were
constructed were used to simulate traditional outside
storage. The farmer was also asked to provide a room
in the house where some bagged maize was stored. An
additional storage crib for cobs was included. Eight
bags of untreated shelled maize were purchased from
farmers at prevailing market price. Admixing of the
pesticide with shelled maize was done using the
fuffle (a device which allows grain to pass through
metal baffles achieving thorough mixing within a few
seconds). Two bags had their maize treated with a
cocktail of 1.6% pirimiphos methyl and 0.3%
permethrin (actellic super dust) at the recommended
dose rate of 50 g/90 kg and then stored in one of the
outside cribs. The other two bags were left untreated
 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
and placed in the other crib to serve as the control. The
remaining four bags were placed on wooden logs in a
room in the house to avoid contact with bare ground.
2.4 Sampling
Sampling from the stores of the six farmers and the
simulation trial was done every 4 weeks. Ten
kilogram sample were removed from each bag and a 1
kg sample taken from it for analysis. Samples were
sieved to remove insects and foreign matter. Moisture
content was determined using a Dicky John meter
(Dickey John Corporation, Auburn. IL, USA).
Samples were then reduced using a riffle divider until
four sub-samples of about 65 g were obtained. Three
sub-samples were analysed and grain sorted into the
following categories: undamaged, holed, tunnelled,
externally damaged, discoloured and broken pieces.
The grains in each category were weighed and their
numbers recorded. The weights were then expressed
as a fraction of the 1 kg sample and converted to
percentage. The grain weight loss was determined
using the count and weigh method as described by
Harris et al. [16].
In the cob trial in Bungoma, two sampling
approaches were used. During the first three months,
sampling was done from the top layer following the
8-compass directions. A total of 17 cobs were picked,
one from the centre and two along each line, making
sure that one was from the cobs next to the crib wall.
After the third month, half of the cobs were removed
and cob samples taken from the new top layer. In the
subsequent sampling, an estimated third of the cobs
were removed before sampling was done with the
final being drawn from the bottom layer. The cobs
were then shelled and both the weight of cob-cores
and grain taken. A 1 kg sample was taken and analysis
done as described above.
2.5 Data Analysis
The data were analysed using both Excel and
statgraphic software. Analysis of variance (ANOVA)
was used to identify factors mostly affecting stored
351
maize. The multiple range test based on the least
square differences (LSD) was used to separate
treatment means at each sampling period. The
practices that were compared using the LSD were
storage in the crib or in the house with and without
treating grain.
3. Results
3.1 Mean Weight of Foreign Matter
Foreign matter was made up of dust, insect frass
and other inedible material. At the onset, foreign
matter was higher in treated samples for both crib and
in-house trials (Table 1). During the 4th and 8th week,
foreign matter in treated samples remained much
higher than in untreated samples. The situation
changed from the 12th week when untreated maize
had 4.6 g/kg compared to 3.8 g/kg for treated samples
stored in the house. Foreign matter in the crib was 3.8
g/kg (untreated) compared with 2.8 g/kg for treated
maize. The subsequent increase in foreign matter was
gradual in treated grains reaching 6.5 g/kg (crib) and
5.1 g/kg (house) while it rose to a peak of 22 g/kg
(crib) and 20 g/kg (house) at the 20th week for the
untreated samples. A comparison between treated and
untreated samples shows positive significant
differences (P < 0.05) occurred from the 4th week up
to the end of the trial (Table 1). However, no
significant differences were observed between the crib
and in house stored grain over the trial period.
Foreign matter sieved from farmers samples appear
to primarily reflect on insect activity within the grains.
Throughout the trial period, foreign matter varied from
1.3 g/kg to 41 g/kg. Out of the six farmers samples
only one deliberately used dust and other maize sieving
to protect grain against storage pests. Subsequent
sampling however, showed no apparent effect since the
number of damaged grains increased. Mean amount of
foreign matter increased from 3.4 g/kg at the beginning
to 19.7 g/kg at the 24th week (Table 2). Amount of
foreign matter from farmer samples poorly compared
with the untreated grain in the simulation trial.
352 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya

Table 1 Changes in weight of foreign matter in simulation trial.

Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 1.3 0.0ab 0.9 0.1b 1.9 0.1b 2.8 0.1a 2.8 0.7a 4.4 2.3a 6.5 0.8a
Crib untreated 0.6 0.4a 0.4 0.0a 0.8 0.8a 3.8 0.2b 7.3 0.6b 13.6 1.7b 22.0 2.7b
House treated 2.1 0.3b 1.2 0.1c 3.4 3.4c 3.8 0.3b 2.6 0.7a 5.5 1.0a 5.1 1.7a
House treated 0.6 0.6a 0.8 0.1b 1.9 1.9b 4.6 0.3b 7.8 0.8b 14.1 0.4b 19.8 2.7b
LSD (0.05) 0.7 0.3 0.7 0.8 2.7 5.9 8.3
Column means followed by the same letter are not significantly different (P < 0.05) (the followings are the same).

Table 2 Weight of foreign matter in individual farmer samples.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 2.7 7.5 - - - -
B 2 8.6 12.2 9.3 - -
C 11.2 21.1 - - - -
D 1.7 7.2 6.8 6.6 23.9 40.7
E 1.5 12.3 26.9 7.5 1.6 2.1
F 1.3 6.8 9.2 19.5 17.2 16.7
Mean 3.4 10.6 13.8 10.7 14.7 19.7

3.2 Mean Moisture Content 3.3 Mean Percent Holed Grain

Moisture content (mc) was adjusted for oven
reading by formula: x + 4.1307/1.2446, generated
from meter calibration trials. It remained steady over
the whole trial period with untreated samples
maintaining a level above the recommended safe
moisture for storage of 13.5% during the first
sampling period (Table 3). The lower moisture content
was due to prevailing sunny weather during the month
of February in the area. A comparison between
treatments shows no significant differences (P < 0.05)
in moisture levels except in the 16th and 24th week.
The moisture content encountered in farmer samples
is a reflection of the varied environment and weather
pattern in the area. At the onset, only one sample had
moisture content above 13.5% considered safe for
storage. Throughout the entire period, moisture figures
fluctuated from 13.7% to 9.5% (Table 4). The average
moisture content varied from 13.7% to 10.1% with the
last reading (24 weeks) indicating some increase.
Similar trend was observed in the simulation trial
indicating no difference in terms of grain moisture for
the different storage environments.
The situation of insect infestation changed from the
12th week when holed grains in untreated samples
increased by a factor of two in crib and 1.5 in the
in-house trial. At the 24th week, holed grains stood at
20% (crib treated) compared with 53% (crib
untreated). In the house treated trial, holed grains were
16% (treated) compared with 56% (untreated) at the
close of the experiment (Table 5). A comparison
between treatments showed no significant differences
(P < 0.05) between crib and house treated grains.
Apparently, the treatments were significantly different
except at the 8th week over the entire trial period.
The level of damage in individual samples varied
from 0.7% to 47% with most having 8%-25%. On the
average, percent holed grains varied from 5.4% to
27.3% but without a definite trend, probably due to
reduction in the number of samples used (Table 6).
From the overall means, the most remarkable increase
in percent holed grain was 20% between 16 and 20
weeks while the least was 1.3% between 20 and 24
weeks. Farmer samples compared more favourably
with the treated grains than the control.
 Assessment of Grain Damage and Weight Loss on Farm Stored 353
Maize in Highlands Areas of Bungoma District, Kenya

Table 3 Changes in moisture content in simulation trial.

Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 13.2 0.0b 13.2 0.2a 12.1 0.2a 12.1 0.0a 11.6 0.2b 12.2 0.0a 12.7 0.1c
Crib untreated 14.0 0.0d 13.2 0.0a 12.4 0.2a 12.1 0.2a 11.1 0.1a 12.1 0.0a 12.6 0.0bc
House treated 13.0 0.0a 12.8 0.0a 12.7 0.1a 12.6 0.4a 11.7 0.2b 12.1 0.1a 12.4 0.2ab
House untreated 13.8 0.0c 13.6 0.5a 12.8 0.2a 12.0 0.6a 11.3 0.1ab 12.2 0.1a 12.2 0.0a
LSD (0.05) 0.1 1.0 0.7 0.8 0.3 0.2 0.3
Column means followed by the same letter are not significantly different (P < 0.05).

Table 4 Moisture content in individual farmer samples.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 12.2 10.2 - - - -
B 11.5 10.2 10.5 9.5 - -
C 12.9 10.6 - - - -
D 13.7 11.2 10.6 9.6 10.5 11.2
E 13.1 10.6 11 11.3 10.7 11.3
F 12.5 10.6 10.3 9.8 10.5 10.7
Mean 12.7 10.6 10.6 10.1 10.6 11.1

Table 5 Changes in percent holed grains in simulation trial.

Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 3.5 0.2b 5.3 0.5b 7.7 0.3a 7.7 0.8a 13.2 3.4a 12.6 6.6a 19.9 6.0a
Crib uneated 0.9 0.4a 0.9 0.1a 3.1 0.8a 14.5 1.0b 43.6 7.3b 42.5 0.5b 52.9 3.6b
House treated 2.7 0.3b 5.1 0.7b 8.7 2.5a 12.0 0.9ab 12.7 1.3a 15.6 2.1a 15.7 4.5a
House untreated 0.3 0.1a 1.5 0.2a 6.0 2.2a 17.4 2.7b 51.5 7.8b 46.6 9.0b 55.8 5.0b
LSD (0.05) 2.7 2.5 2.7 3.3 1.3 1.7 3.2
Column means followed by the same letter are not significantly different (P < 0.05).

Table 6 Mean percentage holed grains in farmer samples.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 4 21 - - - -
B 14 39 4.1 47 - -
C 16 9 - - - -
D 8.3 19 2.6 25 26 21
E 0.7 19 5.4 8.3 10 9.9
F 5.2 7.3 17.0 29 12 21
Mean 5.4 19.1 7.3 27.3 16.0 17.3

By combining the results of the tree environments,
Fig. 1 shows that mean weight of holed grains was the
highest in the crib compared to house and farmer
environment. Individual farmer mean was the lowest
within the three environments.
3.4 Percent Weight Loss
The untreated samples had a gradual increase of
per cent weight loss except the 20th week when it
dropped before it reached a peak of 13.4% by the end
of the trial. Looking at treatment figures, one cannot
fail to notice the apparent influence from the
in-house environment on weight loss. In general,
figures were higher in the in-house trial than in the
crib indicating favourable conditions for the
development of weevils in the former than the latter
354 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya

as observed in Table 7. the other two environments. Considering that the

Table 8 shows that individual farmer samples
suffered a loss of between -0.1% and 12.2%. In most
cases, figures between sampling intervals indicate a
general increase with the average varying from 0.5% to
6% by the 24th week. Mean weight loss in the three
environments (Fig. 2) shows a similar trend as that
experienced in holed grains at 24th week except that
initially the loss for individual farmers was higher than
farmer grain was treated with an insecticide, it was
rather strange that by the end of the six months, farmer
samples compared with crib untreated grain in the
simulation trial. A saving that would have been
expected to occur due to this intervention measure was
hence not achieved. In monetary terms, farmers
experience huge losses besides the accompanying
nutritional loss.

70
60
50
% Holed grains

40
30
20
10
0
4 8 12 16 20 24
Sam pling pe riod (we e k s)

Crib M e an H ous e M e an Farme rs

Fig. 1 Percent weight of holed grains in three environments.

Table 7 Percentage weight loss in maize samples.

Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 0.4 0.2 0.9 0.7 1.2 2.5 2.3
Crib untreated 0.0 0.3 0.4 0.8 3.2 5.3 5.0
House treated 0.4 0.6 0.8 3.3 2.0 2.2 2.2
House untreated 0.3 0.4 -0.1 1.5 6.5 4.1 13.4

Table 8 Percent weight loss on individual farmer samples.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 1.7 2.1 - - - -
B 0.1 5.7 6.3 3.9 - -
C 1.0 2.4 - - - -
D -0.1 1.9 5.0 3.9 5.6 5.3
E 0.2 4.5 -0.1 0.4 1.8 0.4
F 0.04 1.4 1.9 6.1 1.1 12.2
Mean 0.5 3.0 3.3 4.0 2.8 6.0
 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
14
12
10
% wt. loss
8
0
4 8
S ampling period (week s )
C ri b Me a n
Fig. 2 Percent weigh loss in three environments.
3.5 Cumulative Weight Loss
Cumulative weight loss gives a picture of how the
weight loss progressed over the entire trial period.
From a small weight loss of 0.5%, treated grains from
the crib finally recorded 8% by the 24th week (Table
9). The untreated samples started with 0.4% and ended
with 15%, almost double of the treated samples.
Treated maize from the in-house trial had between
0.9% and 11% while weight loss in the untreated
samples increased from 0.4% to 26%, respectively.
The resultant cumulative weight loss is shown in
Table 10. Cumulative weight loss can be viewed in
terms of different backgrounds. As a reflection of the
farmers declining mean stocks, cumulative weight
loss varied from 0.04% at the onset to 3.74% at the
end of trials resulting into lower mean loss compared
to crib and house environments (Fig. 3). The
implication is that when farmers had plenty, a small
loss of 0.04% is inconsequential. However, as the
quantity declined to 390 kg, a weight loss of 3.7%
would have an impact.
3.6 Changes in the Number of Live Insects in Maize
Samples
There was very low infestation at the start of the
trial as shown in Table 11. The situation remained
355
12 16 20 24
H o u s e Me a n Fa rm e rs
almost the same with pest population varying from
0-0.5/kg and 2-9 for treated maize in both crib and
in-house trials between 4 to 12 weeks (Table 11). At
the same time pest numbers had increased in the
untreated samples to 67/kg (crib) and 46/kg (house).
Between 16-24 weeks, pest population increased to an
average of 13/kg (crib treated) compared to 15/kg
(house treated). Pest population in untreated samples
averaged at 67/kg (crib) and 92/kg (house). The
highest live pests of 150/kg were recorded in the
house trial while the peak for the crib was 113/kg.
Apart from untreated samples from the crib, the others
had lower numbers at 24 weeks than the preceding
one. No significant differences were observed
between untreated grain except at 16th and 20th weeks
hence indicating no undue influence from the different
premises used for storage. The most abundant insects
in descending order were: Sitophilus spp, Tribolium
castaneum Herbst and Sitotroga spp..
For the first two sampling interval (4-8 weeks), live
insect population appear to increase except in two
farmer stores. While some farmers had very low
infestation, between 0 and 2 kg, two of them had
populations of above 150 kg. As Table 12 shows, live
infestation appears to decrease from the 12th week
probably due to a range of interventions put in place
356 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya

Table 9 Cumulative percentage weight loss in simulation trial.

Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 0.4 0.5 1.4 2.1 3.3 5.7 8.0
Crib untreated 0.0 0.4 0.8 1.6 4.8 10.0 15.0
House treated 0.4 0.9 1.7 5.0 7.0 9.2 11.4
House untreated 0.3 0.7 0.6 2.1 8.6 12.7 26.1
Average cumulative weight loss of untreated grain in both environments -20.6%;
Average cumulative weight loss of treated grain in both environments -9.7%.

Table 10 Cumulative weight loss calculation on farmers maize stocks.

Storage period Actual stated Quantity put to Mean Cumulative
Mean stocks(kg) Quantity removed
(weeks) consumption other uses % wt loss % wt loss
4 1,135 243 114 120 0.5 0.5
8 901 246.7 106 140 3.0 3.5
12 654.3 174.8 71.2 103.6 3.3 6.8
16 479.5 -55 74 19 4.0 10.8
20 485 95 106.7 -11 2.8 12.8
24 390 390 80.7 309.3 6.0 16.8

60

50
Cumulative % wt. loss

40

30

20

10

0
4 8 12 16 20 24
Sampling period (weeks)

Crib Mean House Mean Farmers

Fig. 3 Cumulative percent weight loss in three environments.
Table 11 Number of live and (total) insects in simulation trial.
Sampling period (weeks)
Treatments 0 4 8 12 16 20 24
Crib treated 1.0 0.0a 0.5 0.5a 0.0 0.0a 0.0 0.0a 11.0 1.0a 17.0 17.0a 12.6 6.5a
Crib untreated 0.0 0.0a 0.5 0.5a 16.5 7.5ab 65.5 60.5a 59.5 1.5a 28.0 0.0a 112.5 34.5b
House treated 1.0 1.0a 8.5 1.5a 2.0 2.0a 2.0 1.0a 8.5 6.5a 23.5 4.5a 15.0 1.0a
House untreated 0.0 0.5a 4.0 3.0ab 45.5 18.5b 31.5 0.5a 80.5 5.5c 150.5 51.5b 43.3 5.5ab
LSD (0.05) 22 6.7 39.4 118.8 17.1 106.8 69.8
Column means followed by the same letter are not significantly different (P < 0.05).
 Assessment of Grain Damage and Weight Loss on Farm Stored 357
Maize in Highlands Areas of Bungoma District, Kenya

Table 12 Mean number of live insects farmer samples.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 1 (8) 2 (178) - - - -
B 15 (16) 175 (364) 4 (312) 24 (204) - -
C 22 (53) 0 (122) - - - -
D 2 (4) 46 (153) 14 (36) 1 (71) 0 (1,128) 0 (590)
E 32 (34) 438 (450) 1 (38) 0 (132) 3 (17) 0 (80)
F 0 0 (44) 0 (104) 0 (420) 1 (268) 1 (676)
Mean 12 (19.2) 109 (219) 5 (123) 6 (207) 1 (471) 0 (449)

by farmers. This is also reflected in the average
number of live insect pests which reached a peak of
109/kg during the 8th week. By subtracting live
insects from the total in bracket, it is easy to
compare the effectiveness of farmers pest control
options. Between 4 and 24 weeks, dead insects rose
from 4 to 1,128 kg sample. The rate of infestation in
farmer samples from the 12th week compared very
well with the treated grains in the simulation trial
(Table 11)
3.7 Changes in Various Parameters of Cob Maize
Stored in Crib
grain discolouration declined from 1.1% to zero by the
end of the trial, indicating no contribution by storage
period or pests.
3.7.4 Percent Weight Loss
A modest weight loss figures of between 0.6% and
2.6% was recorded between 0-8 weeks. Increased pest
activity in the latter three sampling intervals is
responsible for higher weight loss of up to 12.8% by
the 12th week. The resultant cumulative weight loss
reached 22.2% (Table 13) which is close to 25.8%
recorded in the untreated maize stored in the house.
3.8 Condition of Farmer Maize Stocks

3.7.1 Percent Moisture Content
At the onset, the moisture content was at a high of
15.3%. As storage progressed, grain moisture
dropped to 9.3% due to continuous hot weather
(Table 13). This contrasted with what happened in
the 24th week when moisture increased to 13.5%.
The rise could be attributed to increased pest activity
or equilibration to conform to the surrounding
environment.
3.7.2 Percent Weight of Holed Grains
Table 13 shows that sampled cobs had no pest
damage at the onset of the trial, however the situation
changed and holed grains rose from 1.4% to 7.5%
between 4 to 8 weeks. Holed grains progressively
increased to a maximum of 26.2% by the 24th week.
3.7.3 Percent Discoloured Grains
Although the cobs sampled at the onset had no
discoloured grains, the subsequent ones declined from
2.2 to 1.2 by the 8th week. In the last three samplings,
Contrary to the belief that Bungoma was endowed
with high potential, the sample of six farmers
indicates that availability of maize was not always
guaranteed. The issue was not also restricted to
farmers who do not harvest large quantities of over 20
bags of 90 kg as shown in Table 14. About half of the
farmers had cleared their maize stocks by the 20th
week. On average, maize stocks at the onset were
1,135 kg and the quantity dropped gradually to 390 kg,
equivalent of 13 to 4 bags by the end of the trial.
Household consumption pattern indicate wide
variation from a low of 16.8 kg to 176 kg per month
(Table 15). Average consumption fluctuated over the
entire storage period with the highest amount of 114
kg and the lowest of 71 kg per month. Pro-rata decline
was not established implying that farmers used more
maize at different times. By reconciling quantities
consumed and the mean stocks, it can be seen that
farmers also use their maize for other purposes (Table 16)
358 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya

Table 13 Changes in various parameters of maize stored on cob in the crib.

Period Moisture Weight holed Wt. discoloured Number Cumulative %
Bulk density Weight frass % wt loss
(weeks) content grains grains insects wt loss
0 342 15.3 0.1 0.0 0 0 0.6 0.6
4 341.1 13.1 0.5 1.4 2.2 5 0.0 0.6
8 325.7 11.1 2.8 7.5 1.2 84 1.0 1.6
12 335.7 10.3 2.3 3.8 1.8 76 2.6 4.1
16 320.7 9.3 3.8 11.2 0 97 1.3 5.4
20 317.3 10.9 6.4 26.2 1.1 152 3.8 9.2
24 297 13.5 14.5 20.3 0.5 193 12.8 22.2

Table 14 Decreasing maize stocks (kg) in farmers stores.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 1,980 1,800 - - - -
B 348 225 187 28 - -
C 477 42 - - - -
D 630 279 225 135 105 62
E 405 360 315 225 180 74
F 2,970 2,700 1,890 1,530 1,170 1,034
Mean 1,135 901 654.3 479.5 485 390

Table 15 Consumption pattern (kg).

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 140 140 - - - -
B 92 140 16.8 80 - -
C 92 92 - - - -
D 92 96 62 80 88 70
E 92 84 100 52 116 88
F 176 84 106 84 116 88
Mean 114 106 71.2 74 1.607 80.7

Table 16 Decreasing maize stocks after consumption pattern.

Sampling period (weeks)
Farmer code 4 8 12 16 20 24
A 1,980 1,840 (1,800) - - - -
B 348 208 (225) 191.2 (187) 111.2 (28) - -
C 477 385 (42) - - - -
D 630 534 (279) 472 (225) 392 (135) 304 (105) 234 (620)
E 405 321 (360) 221 (315) 169 (225) 53 (180) -34 (74)
F 2,970 2,886 (2,700) 2,780 (1890) 2,696 (1,530) 2,580 (1,170) 22,492 (1,034)
Figures in parenthesis denote the actual quantities of maize during sampling.

due to the fact that maize is an important cash crop in maize both as a staple food and cash crop [17]. Due to
the area. the prevailing weather conditions of a mean rainfall of
1,899 mm and temperatures of 22 C, Bungoma had
4. Discussion
only one maize crop in a year [18]. For this reason it
The area where the study was undertaken relies on was important for farmers to store their maize grain
 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
for almost a year hence predisposing the crop to insect
infestation. This made it prudent for them to preserve
and protect the harvested maize. It was also observed
that maize, or its derivative was cooked on daily basis
save for an occasional meal when other alternatives
were used. The importance of maize to the community
therefore underscores the need to monitor the quality
of the grain in farmer stores.
Consumption patterns established indicate that
Bungoma farmers reserved 13 bags at the start of the
storage season later declining to 4 bags (Table 15). An
average of 120 kilograms of maize was put to other
uses. In an earlier study [19], Kitui farmers appeared
to be more food secure with an average of 21 bags of
maize at harvest time. It can be argued that farmers in
food deficit areas have evolved clandestine ways of
conserving their grain to avoid famine.
Storage insect pests were the main causes of food
loss in the district as has been found in other countries
[20]. The main pests were the maize weevil, Sitophilus
zeamais, Angoumois grain moth, Sitotroga cerealella,
rust red flour beetle, Tribolium castaneum and other
secondary pests of minor importance. As simulation
trial showed, there was low initial infestation but the
numbers increased as the storage period progressed
(Table 11). The variance between the number of
insects in the treated and untreated maize was an
indication of the extent of efficacy of the product
applied to control the storage pests. As shown in
Table 12, the high numbers of dead insects showed
some farmers were able to reduce infestation probably
to manageable levels. A corresponding damage of
16%-20% of treated grain in the simulation trial
(Table 5) was a clear indication of failure of the test
chemical since previous studies [21] showed that the
same product was able to maintain damage levels of
5%. These observed high damage levels could have
resulted from delayed treatment and possibly lowered
efficacy of the applied insecticide.
Damage caused by the primary pests becomes
visible after adults emerge through the exit holes and
359
by this time loss in grain quality, food value and
weight, has already taken place. The real impact of
pest infestation could be understood by looking at the
consequent accumulative weight loss in both
simulation trial and farmer stocks if weight loss was
related to the declining maize stocks. Average
cumulative weight loss of 20.6% and 9.7% for treated
and untreated maize respectively was realised in
Bungoma (Table 9). Weight loss experienced in
farmer stocks in Bungoma, compared well with
untreated samples. Compared to individual farmer
stocks, simulation trials also experienced higher losses
suggesting that delayed grain treatments (maize used
in simulation trial had been stored for 3-4 months after
harvest) increased chances of attack due to pressure of
emerging insects irrespective of intervention methods
employed. This supports the recommendation of shell,
treat and store campaign during post harvest
dissemination by Ministry of Agriculture and Kenya
Agricultural Research Institute staff without which
achievement in eradicating storage pests in stored
grain at farm level would be futile. From the above
loss figures, it was apparent that farmers in Bungoma
managed to a certain extent to reduce grain damage
and hence weight loss by the sixth month through
grain treatment thereby experiencing marginal saving.
Weight loss of 12.2% on individual farmer samples
at the 24th week disagreed with studies conducted in
Tanzania where after taking into account removals for
home consumption, the real food losses were less than
2% over a period of 7-9 months [22]. An earlier study
in Kitui was in agreement to this study [19] and hence
suggested that Bungoma farmers were ineffective in
their control measures.
The consequence of insect pest infestation is a
reduction in grain weight, which ideally should be
reflected in price difference between good and
damaged grain. However, the price depends on the
level of damage and availability of grain in local
markets, with the one with extensive damage being
liable to rejection by consumers. Weight loss is
360 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
therefore a very important indicator of quality
deterioration.
Based on these weight loss results, with an average
production figures 1.7 million per 90 kg bag (153,000
tons) in Bungoma (Ministry of Agriculture, 2000,
2001), a loss of 340,000 90 kg bags (20%) would
occur within 6 months of storage period. At
approximately Kshs 1,500 per 90 kg bag, a 510
million Kenya Shillings (6.3 million US dollars) in
total would be lost. Such an amount could, if losses
were averted, be used to mount various projects
including training to ensure food security.
4.1 Coping Strategies
Every pest control operation is aimed at eradicating
the existing population. Farmers, in their effort to
achieve their pest control goal, have to contend with a
variety of problems ranging from expensive and often
unavailable or expired and adulterated pesticides [23]
and of recent times, increased tolerance by important
storage pests. These are some of the points that must
be borne in mind when trying to judge the success of
farmers methods particularly during on-farm surveys
and trials. In an attempt to maintain grain quality,
some farmers are known to use different methods to
control pests. While some rely on the recommended
chemical dusts, others use ash mixed with such
chemicals in the hope of increasing the potent of the
former. When the method adopted appear to fail,
farmers make quick decision to dispose the grain
either through sale, feed livestock, pay labour or
mortgage, it against a crop failure in the near future.
All these appear to mitigate against heavy damage
hence weight loss.
Farmers in Bungoma coped with insect infestation
by treating grain with Actellic super and Malathion
dusts. Others due to financial constraints used maize
debris to protect their grain ending up with disastrous
results and finally resorting to chemical use when it
was too late. On the positive side however, farmers
were fully aware on the need to protect their stored
maize irrespective of the various constraints
associated with rural farming that impedes them to
utilise the acquired knowledge.
5. Conclusion
The trials showed that farmers are continually
losing a lot of stored grain to insects after harvest due
to delayed treatment, non-treatment, incorrect dosage
and lack of re-treatment especially for those storing
for more than 6 months. Use of adulterated and
non-recommended chemicals also contributed to
storage losses. However a few of the farmers who
followed good grain storage practices managed to
maintain pest free quality grain even for more than 6
months of storage. Strong efforts must be made to
bring insect damage level to less than 5% of stored
maize if farmers have to reap benefit from their sweat
during grain production.
To achieve this noble task, great awareness should
be directed to the farmer to enable them gain the
necessary knowledge, for example, timely harvesting,
judicious use of chemicals, store hygiene, etc. which
will translate into well maintained quality crop. In
addition, development of IPM strategies for control of
these pests will further lead to improvement of stored
grain and consequently, rural food security will be
enhanced. It is important to note that while the trial
borrowed greatly from farmers practices, the results
reflected the true situation on the ground and can be
used as a reference for further work to determine
weight losses as a prelude for the development of
intervention strategies.
Acknowledgments
Profound thanks to KARI, Ministry of Agriculture
and DAO Bungoma for their support during the
execution of the trial. Further appreciation is extended
to Francis Mwangi for technical support during grain
sampling and analysis. More thanks and appreciation
also go to Prof. James Harper for the valuable
comments and support. The project was a joint effort
 Assessment of Grain Damage and Weight Loss on Farm Stored
Maize in Highlands Areas of Bungoma District, Kenya
between KARI and North Carolina State University
(NCSU) with part sponsorship from NCSU Department
of Entomology and Blanton J. Whitmire Developing
Countries Integrated Pest Management and Post
Harvest Food preservation Program Endowment.
References
[1] Anon, Economic survey, Kenya National Bureau of
Statistics, Ministry of Planning, 2011, pp. 151-167.
[2] N. Raja, S. Albert, S. Ignacimuthu, S. Dorn, Effect of
plant volatile oils constituents towards stored cowpwa
Vigna unguiculata (L.) Walpers against Callosobruchus
maculatus (F.) (Coleopters: Bruchidae) infestation,
Journal of Stored Products Research 37 (2001) 127-132.
[3] D.P. Papachristos, D.C. Stampoulos, Repellent, toxic and
reproduction inhibitory effects of essential oil vapours on
Aconthoscelids obtectus (Say) (Coleoptera: Bruchidae),
Journal of Stored Products Research 38 (2002) 117-128.
[4] S. Singamony, I. Anees, T. Chandrakala, Z. Osman,
Efficacy of certain indigenous plants products as grain
protectants against S. oryzae (L.) and R. dominica (F),
Journal of Stored Products Research 22 (10) (1985) 21-23.
[5] L.E. Jakai, R.A. Daoust, Insect pests of cowpeas, Annual
Review of Entomology 31 (1986) 95-119.
[6] A. Jilani, C. Saxena, Repellents and feeding deterrent
effects of tumeric oil sweet flag oil and neem based
insecticides against lesser grain borer (Col: Bostrichidae),
Journal of Economic Entomology 83 (2) (1990) 629-634.
[7] F. Abebe, T. Tefera, S. Mugo, Y. Beyene, S. Vidal,
Resistance of maize varieties to the maize weevil
Sitophilus zeamais (Motsch.) (Coleoptera:Curculionidae),
Afri. J. Biotechnol. 8 (2009) 5937-5943.
[8] C.P.F. De Lima, The assessment of losses due to insects
and rodents in maize stored for subsistence in Kenya,
Tropical Products Information 38 (1979) 21-26.
[9] P. Golob, P. Changjareon, A. Ahmed, J. Cox,
Susceptibility of Prostephanus truncatus (Horn) to
insecticides, Journal of Stored Products Research 21 (3)
(1985) 141-150.
[10] C. Hencles, Investigations into Insect Dynamics, Damage
and Losses of Stored Maize. An Approach to IPM on
Small Farms in Tanzania with Special Reference to
Prostephanus truncatus (Horn) (Coleoptera: Bostrichidae)
Hamburg, Germany GTZ., 1982, p. 124.
[11] P.A. Tyler, R.A. Boxall, Post-harvest loss reduction
programme: A decade of activities: What consequences?,
Tropical Stored Product Information 50 (1984) 4-13.
361
[12] R.A. Boxall, Damage and loss caused by the larger grain
borer Prostephanus truncates, Integr. Pest Manage. Rev.
(7) (2002) 105-121.
[13] P. Golob, R. Hodges, A study of the outbreak of
Prostephanus truncatus (Horn) in Tanzania, Tropical
Information Report No. G. 164 (1982) 23.
[14] S.K. Muhihu, G.N. Kibata, Developing a control
programme to combat an outbreak of Prostephanus
truncatus (Horn) (Coleoptera: Bostrichidae) in Kenya,
Tropical Science 25 (1985) 239-248.
[15] R. Jaetzold, H. Schimdt, Farm Management Handbook of
Kenya. Vol. II, Parts A, B and C, Published by the
Ministry of Agriculture, Kenya in Co-operation with the
German Agricultural Team (GAT) of the German Agency
for Technical Co-operation (GTZ), 1983, pp. 309-356.
[16] K.L. Harris, C.J. Lindblad, Postharvest Grain Loss
Assessment Methods, American Association of Grain
Cereals, 1978, pp. 83-91.
[17] Anon Economic Review of Agriculture, Ministry of
Agriculture, 2011, p. 45.
[18] K. Mutambuki, C.M. Ngatia, J.N. Mbugua, Post Harvest
Loss Assessment in Bungoma and Kitui districts, Project
number KARI CS/SP/01/1, 2001, pp. 1-53.
[19] K. Mutambuki, C.M. Ngatia, Loss assessment of on-farm
stored maize in semi arid areas of Kitui District, in: I.
Lorini, B. Bacultchuk, H. Beckel, D. Deckles, et al.
(Eds.), Proceedings of the 19th International Working
Conference on Stored Product Protection Campinas, Sao
Paulo, Brazil, 2006, pp. 15-23.
[20] P. Golob, C. Moss, M. Dales, J. Fidgen, I. Gudrups, The
Use of Spices and Medicinals as Bioactive Protectants for
Grains, FAO Agricultural services Bulletin 137, FAO,
Rome, Italy, 1999.
[21] K. Mutambuki, C.M. Ngatia, Efficacy and stability of
pirimiphos methyl/permethrin (actellic super) formulated
with organic filler, in: Kimondo Mutambuki (Ed.),
Technical reports on pesticide screening against stored
pests from 1995-2007, KARI/NARL CS/SP/2007.
[22] P. Golob, Current status of the larger grain borer
Prostephanus truncatus (Horn) in Africa, Insect Science
and Its Application 96 (1988) 737-745.
[23] P. Golob, N. Marsland, B. Nyambo, K. Mutambuki, A.
Moshy, E.C. Kasalile, et al., Coping strategies employed
by farmers against the larger grain borer in East Africa, in:
Jin Zuxun, L. Yongstang, T. Xianchang, G. Liangua
(Eds.), Proceedings of 7th International Working
Conference on Stored Product protection, Beijing, 1998,
pp. 1772-1781.
Journal of Agricultural Science and Technology B 2 (2012) 362-367
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D
DAVID PUBLISHING

Molecular Characterization of Cowpea Breeding Lines

for Striga Resistance Using SCAR Markers

Lucky O. Omoigui1, Mary Yeye2, Boukar Ousmane3, 4, Bhavani S. Gowda5 and Michael P. Timko5
1. Department of Plant Breeding & Seed Science, College of Agronomy, University of Agriculture, P.M.B. 2373, Makurdi, Nigeria
2. Department of Plant Science, Ahmadu Bello University, P.M.B. 1044, Zaria, Nigeria
3. International Institute of Tropical Agriculture, P.M.B. 5320, Ibadan, Nigeria
4. C/o Lambourn Ltd, Carolyn House, CR93EE, Croydon, UK
5. Department of Biology, University of Virginia, Charlottesville, Virginia 22904, USA

Received: June 16, 2011 / Published: March 20, 2012.

Abstract: S. gesnerioides (Willd) Vatke is a major biological constraint to cowpea production in the dry savanna of sub-Saharan Africa.
Yield losses caused by S. gesnerioides in these regions are estimated in millions of tons annually, and prevalence of Striga soil
infestation is steadily increasing. The availability of molecular markers tightly linked to S. gesnerioides resistance genes opens up the
possibility of applying Marker-Assisted Selection (MAS) to cowpea and would fast track the process of developing resistance varieties
to the parasite. In the present study, we report the use of Fast Technology for Analysis (FTA) also known as PlantSaver Cards
(Whatman FTA), developed by Flinder Technology associate to retrieve DNA from plant tissue for molecular analysis. A total of 100
F2 individual plants derived from two crosses were validated for SG3 resistance using two different SCAR markers (MahSe2 and C42B)
linked to Striga race 3 (SG3) and 5 (SG5) resistance in other segregating populations. Genomic DNA was successfully recovered from
leaf tissues of cowpea pressed onto FTA classic card and the DNA obtained from the FTA papers was found to be suitable for molecular
analysis by PCR-based techniques. The marker efficiency of SCAR MahSe2 and C42B in detecting SG3 resistance was 98.5% and
93% respectively. This result revealed the utility of SCAR markers in cowpea breeding programme. Therefore, the application of MAS
using FTA technology has the potential to increase efficiency of selection and for molecular characterization of cowpea lines for Striga
resistance..

Key words: Marker assisted selection, S. gesnerioides, race, FAT technology.

1. Introduction identifying resistant and susceptible genotypes. Often

Striga gesnerioides is a major biological constraint
to cowpea production in sub-Saharan Africa. S.
gesnerioides is difficult to control, and once
established few means are available to counter its
impact on yield. Yield losses raging from 83%-100%
have been reported [1, 2]. As a result, development and
deployment of resistance crop varieties remain the
most effective manner to combat the menace presented
by S. gesnerioides. Classical tools have been used in
Corresponding author: Lucky O. Omoigui, Ph.D., research
fields: plant breeding, molecular genetics. E-mail:
lomoigui@yahoo.co.uk.
time this classification is affected by the environment
and requires a long time to develop resistance variety.
Therefore, effective method such as molecular markers
is required to facilitate characterization of cowpea lines
for resistance to Striga, and this would speed-up the
process of crop delivery to farmers. Access to simple,
low cost tools for the molecular study is central to
generate the knowledge required.
A frequent limitation for studying DNA at molecular
level is the ability to obtain high quality DNA from
plant tissues. Plant tissues to be analyzed must be
collected and preserved in order to maintain integrity of
the DNA until they can be processed. This poses
 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR Markers
challenges when sample numbers are large and when
working in the field, most especially in the tropical and
sub-tropical regions where laboratory facilities are
limited. Field and greenhouse studies are thus
constrained by the resources required for sample
collection and transportation, placing restrictions on
the number of samples that can be collected in a given
time and size. Timely processing or storage of the
samples before they spoil can also be difficult in
locations where access to well equipped laboratory
facilities is limited. Intact high molecular weight plant
DNA is essential for molecular studies and genomic
DNA library construction. Previous traditional
isolation methods required grinding plant tissue in
liquid nitrogen and transferring it to a preheated
extraction buffer [3, 4]. Liquid nitrogen can be difficult
to procure in remote locations; thus, a method not
requiring its use would be helpful.
The use of FTA cards (Whatman) provide a
simple alternative method for collection, storage and
retrieval of genomic DNA for molecular study
especially when operating in developing countries and
regions remote from laboratories facilities. The use of
FTA cards have been reported from a wide range of
biological sources like whole blood, tissue, plasmid,
plant materials, and microorganisms, etc [5, 6]. FTA
treated matrix cards are impregnated with a proprietary
mix of chemicals containing strong buffers, free radical
trap and protein denaturants that lyses cell membranes
on contact, physically entrap DNA, and stabilizes and
protects DNA from nuclease, oxidation, UV damage
and from microbial and fungal degradation [7, 8].
FTA cards have also been used in the collection of
DNA from bacterial cells for molecular analysis [9].
They have also been used in the collection of leaf tissue
of a variety of plants, including soybean [10], tomato,
tobacco and grapes [11], molecular genetic analysis of
scleractinian corals in remote environmental locations
[12], sampling, recovery and molecular
characterization of viral pathogens and virus-derived
363
transgenes from plant tissue [11]. It has also proven to
be useful for large-scale plant DNA isolation for use in
marker-assisted selection [13]
To-date limited successes have been recorded to its
potential use in collection of cowpea leaf material for
long time storage and retrieval of total genomic DNA.
Therefore, we assessed the potential of the FTA
matrix card system as an effective technology for
sampling and retrieval of genomic DNA from cowpea
tissue and their subsequent molecular analysis using
SCAR markers.
2. Materials and Methods
2.1 Plant Material
Cowpea (Vigna unguiculata (L.) Walp.) parental
lines and the derived F2 populations were grown under
standard greenhouse condition in a pot culture
inoculated with Striga seeds at IITA, Kano Station,
Nigeria in 2008. Genomic DNA was extracted from
leaf tissue of one-month old plants using the FTA
Plantsaver cards. The second young leaf was excised
from the plant and placed in square of the FTA card.
Leaf sample was covered with a parafilm paper and a
pestle was used to press the leaf sample extracted onto
the FTA paper until both sides of the FTA were
soaked. Paper towel soaked in 70% ethanol was used to
clean pestle in between samples to prevent cross
contamination.
2.2 Preparation of FTA Tissue Print
The young leaf was placed on the FTA Plantsaver
card covered with parafilm paper, and pressure was
gently applied with a pestle briefly until plant material
was sufficiently transferred to the card. The cards
were allowed to dry for one hour and brush off any
plant material with tissue paper. After air-drying,
FTA cards were placed in a paper punch and stored
at ambient temperature in a dry location.
364 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR Markers
2.3 Preparation of Samples for PCR Analysis
A disc from the dried FTA tissue print was removed
using a clean Haris micro punch and placed the disc
directly into a 1.5 mL Eppendorf tube. In between
samples, the Haris micro punch was cleaned with a
tissue dampened with 70% ethanol and taking a disc
from a blank, unused FTA card to prevent cross
contamination. The disc was washed twice with 200 L
of 70% ethanol, incubating for five minutes for each
wash, followed by a repeated wash with 200 L of FTA
reagent incubating for 3 min at room temperature and
the liquid was discarded. The tubes were inverted and
drained on a paper towel and air dried for
approximately 1 h. After drying the disc was
transferred to PCR tube for PCR analysis.
2.4 PCR Analysis
PCR analysis was done with 2 primers MahSe2 and
C42B. The PCR mixture (25 L final volume) contained,
in addition to the purified 2 mm FTA DNA disc
containing the DNA sample, a final concentration of 18
L of sterilized water, 2.5 mM each of DNTPs mix and
10 PCR buffer, 0.05 L of Taq polymerase, and 1 L
of each of the forward and reverse primers (synthesized
by IDT, Coraville, Iowa). PCR reactions were
performed on a heated lid thermal cycle (Biometra)
using the thermal cycle (Biometra) operated as
following conditions: 35 cycles of denaturation at 94 C
for 30 s, followed by annealing at 57.5 C for 30 s and
extension at 72 C for 2 min. A final extension cycle of
10 min at 72 C was added to ensure completion of the
final amplification products. For C42B marker, a similar
procedure was followed but the annealing temperature
was adjusted to used was 67.5 C.
2.5 Analysis of PCR Product
A 25 L of the final PCR product was
electrophoresed on a 3% agarose gel stained with
ethidium bromide. The gels were run for approximately
1 hour 30 minutes at 120 volt in 1 X TAE buffer (45
mmol L-1 glacial acetic acid, 0.5 mmol L-1 EDTA, pH,
8.4). A 1 kb DNA ladder was loaded in the first well for
band size determination of PCR products. The
ethidium bromide-stained gel was visualized on an UV
transilluminator and images photographed using a
polaroid camera.
3. Results and Discussion
Genomic DNA was successfully isolated from leaf
tissue of cowpea pressed onto FTA classic cards.
Plant DNA eluted from FTA cards stored for over
nine months at room temperature was found suitable
for molecular analysis by PCR-based techniques in a
manner equivalent to that offered by traditional
isolation methods (liquid nitrogen). Two primers
MahSe2 and C42B linked to Striga resistance gene in
LG1 were used to discriminate between resistant and
susceptible lines in the F2 populations. DNA of the
resistant lines using markers linked to the race-specific
Striga (SG3) was detected from the population
collected on FTA cards. Unique bands were produced
by the two markers. MahSe2 amplified two fragments
of similar size to the AFLP fragment observed in B301,
one band was present in both the resistant and
susceptible genotypes and a lower fragment that was
polymorphic, being present in resistant genotypes but
absent in susceptible genotypes. On the other hand, the
primer C42B identifies resistant lines with a single
band while susceptible line had no band (Figs. 1-4).
PCR amplification of DNA generated from cowpea
DNA eluted from FTA cards was compared to that of
conventional breeding method. FTA was found to be
suitable for recovery of genomic DNA and for
molecular characterization of the segregating
population for resistance and susceptible to S.
gesnerioides. The marker indicator was quite similar
with the phenotypic classification. Results obtained
from FTA sampled materials were effective and
reproducible in all the three segregating populations
used. The studies described here demonstrate that
FTA offers a simple, sensitive and specific tool
appropriate for molecular characterization of plant
 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR Markers 365

L C P1 P2 1R 2S 3R 4R 5R 6R 7R 8R

450 bp

Resistant Susceptible
Fig. 1 Results of PCR amplification of genomic DNA by C42B marker for the F2 progenies derived from Borno brown
IT03K-338-1. L = 1 kb ladder, C = control without Genomic DNA template. P1 = Borno brown, P2 = IT03K-338-1. R and S
indicate resistant and susceptible, respectively.

L C P11 P2 1R 2R 3R 4R 5R 6R 7R 8S

500bp

Resistant Susceptible
Fig. 2 Results of PCR amplification of genomic DNA by MahSe2 for the F2 progenies derived from Borno brown
IT03K-338-1. L = 1 kb ladder, C = control without genomic DNA template. P1 = Borno brown, P2 = IT03K-338-1. R and S
indicate resistant and susceptible, respectively.

Fig. 3 Agarose gel electrophoretic analysis of PCR amplified product using MahSe2 for the F2 progenies derived from
Borno brown x IT97K-499-35. L = 1 kb ladder, C = control without genomic DNA template. P1 = Borno brown, P2 =
IT97K-499-35, line 1-6 are resistant with double bands while line 8 with single band is susceptible. R and S indicate resistant
and susceptible.
366 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR Markers
Fig. 4 Agarose gel electrophoretic analysis of PCR
amplified product using C42B marker for the F2 progenies
derived from Borno brown x IT97K-499-35. L = 1 kb
ladder, C = control without genomic DNA template. P1 =
Borno brown, P2 = IT97K-499-35, line 1-6 are resistant
with double bands while line 8 with single band is
susceptible. R and S indicate resistant and susceptible.
genomic DNA isolated from cowpea leaf tissues. We
believe that the application of this technology has the
potential to significantly enhanced cowpea breeding
program and efficiency of breeder to speed up the
process of developing and deploying cowpea Striga
resistance varieties to farmers. There was a positive
significant correlation (0.864) between the two makers
and similarity with that of the phenotypic classification
(Table 1). This indicates that these markers are
efficient in detecting resistance and susceptibility in
segregating population and that the marker techniques
are valuable.
FTA technology was also used to sample F3
cowpea plants for molecular characterization using
BIONEER AccuPower PCR PreMix (AccuPower
PCR PreMix is a new, and ready-to use PCR reagent
optimized PCR amplifications). The PCR
amplification also produced uniform banding patterns
on 3% agarose gel, indicating complete amplification
of DNA samples, regardless of the method used. The
AccuPower PCR PreMix yielded the expected result
in the F3 populations tested with primer pair MahSe2.
The results obtained here indicate that it is possible
to use this effective technology for sampling and
retrieval of DNA from plant tissue and their subsequent
molecular analysis for molecular characterization in
Table 1 Spearman rank correlation coefficient of marker
similarity matrix.
Phenotype
MahSe2 0.8346 (< 0.0001)
Borno brown x IT97K-499-35
C42B 0.8382 (< 0.0001)
MahSe2 0.8207 (< 0.0001)
Borno brown x B301
C42B 0.91437 (< 0.0001)
MahSe2 0.8570 (< 0.0001)
Borno brown x IT03k-338-1
C42B 0.8847 (< 0.0001)
Correlation values above 0.8 are considered good association
[14]. Values in parentheses are level of significance expressed as
probability.
cowpea for identifying resistance to S. gesnerioides.
Similar results have been reported in other crop using
this procedure for marker genotyping [15, 16]. The
ability to obtain and store the prints at ambient
temperatures means that these tests could be employed
for wide-scale studies in the field to enhance cowpea
breeding programme for researching cowpea for S.
gesnerioides resistance. The benefits of this technology
have important implications for improving the
efficiency of molecular characterization of cowpea
genotypes for resistance to Striga in the laboratory
especially when working in remote area and in
developing countries where access to laboratory
facilities, chemicals and equipment are limited. Results
obtained from FTA sampled materials were effective
and reproducible in our hands from the three
populations used. The studies described here
demonstrate that FTA offers a simple, sensitive and
specific tool appropriate for molecular characterization
of plant genomic DNA isolated from plant tissues.
4. Conclusion
Results of this study demonstrated that FTA is
effective and sensitive method for sampling, storage
and retrieval of genomic DNA from cowpea leaf
tissues. The important advantage brought by FTA
technology is the ability to fix and reliably preserve
DNA from plant tissue. Benefits of this technology are
realized at both the sampling and processing phases.
Sampling plant material with FTA cards is reduced to
sample and is thus rapid and uncomplicated. The
 Molecular Characterization of Cowpea Breeding Lines for Striga Resistance Using SCAR Markers
ability to store pressed and fixed samples for a long
time at ambient temperatures also significantly reduces
concerns regarding DNA degradation during storage.
The potential number of samples that can be collected
within a given time and location is significantly
increased compared to traditional methods of using
liquid nitrogen, and one no longer needs to depend on
storage of rapidly frozen materials for DNA Isolation.
Application of this technology has the potential to
significantly increase the ability of breeders to bring
modern analytical techniques to bear in an effort to
introgress durable and stable resistance to S.
gesnerioides into local germplasm and speed up the
process of developing and deploying improved Striga
resistant cowpea varieties to farmers.
Acknowledgments
The authors thank personnel of the Timko lab at the
University of Virginia, USA, for technical support and
advice. This work was supported by a grant from the
Kirkhouse Trust Foundation to MPT.
References
[1] V.D. Aggarwal, Cowpea Striga research, in: S.R. Singh,
K.O. Rachie (Eds.), Cowpea Research Production and
Utilization, John Wiley and Sons, 1985, pp. 335-340.
[2] A.M. Emechebe, B.B. Singh, O.I. Leleji, I.D.K. Atokple,
J.K. Adu, Cowpea Striga problems and research in Nigeria,
in: S.K. Kim (Ed.), Combating Striga in Africa,
Proceedings of an International Workshops organized by
IITA, ICRISAT and IDRC, IITA, Ibadan, Nigeria, Aug.
22-24, 1991, pp. 18-28.
[3] S.L. Dellaporta, J. Wood, L. Hicks, A plant DNA
minipreparation: Version 11, Plant Molecular Biology
Reporter 1 (1983) 19-21.
[4] T. Mohapatra, R.P. Sharma, V.L. Chopra, Cloning and use
of low copy sequence genomic DNA for RFLP analysis of
somaclones in mustard [Brassica juncea (L.) Czern and
Coss], Science 62 (1992) 482-484.
367
[5] K.A. Lampel, D. Dyer, L. Komegay, P.A. Orlandi,
Detection of bacillus spores using PCR and FTA filters,
Journal Food Protection 67 (2004) 1036-1038.
[6] Whatman Inc. FTA protocols collect, transport, archive
and access nucleic acids-all at room temperature, Journal
Virol (2002) 1-4.
[7] S. Kubo, Y. Fujita, Application of FTA technology to
extraction of sperm DNA from mixed body fluids
containing semen, Legal. Med. 8 (2006) 43-47.
[8] L.M. Smith, L.A. Burgoyne, Collection, archiving and
processing DNA from wildlife samples using FTA data
basing paper, BMC Ecology 4 (2004) 4.
[9] D. Rajendram, R. Ayenza, F.M. Holder, B. Moran, T.
Long, H.N. Shah, Long-term storage and safe retrieval of
DNA from microorganisms for molecular analysis using
FTA matrix, J. Microbiologycal Methods (2006) 582-592
[10] J. Lin, R. Fleming, J. Kuo, B.F. Mathews, J.A. Saunders,
Detection of plant genes using a rapid, nonorganic DNA
purification method, Biotechechnology 28 (2000) 346-350.
[11] J. Ndunguru, N.J. Taylor, J. Yadav, H. Aly, J.P. Legg, T.
Aveling, et al., Application of FTA technology for
sampling, recovery and molecular characterization of viral
pathogens and virus-derived transgenes from plant tissues,
Journal of Virology 2 (2005) 45.
[12] M.J. Crabbe, A novel method for the transport and
analysis of genetic material from polyps and zooxanthellae
of scleractinian corals, Journal Biochemistry and
Biophysics 57 (2003) 171-176.
[13] M.N. Mbogori, M. Kimani, A. Kuria, M. Lagat, J.W.
Danson, Optimization of FTA technology for large scale
plant DNA isolation for use in marker assisted selection,
African Journal of Biotechnology 5 (2006) 693-696.
[14] F.J. Rohlf, NTSYS.PC Numerical Taxonomy and
Multivariate Analysis System, Version 1.8, Applied
Biostatistics Inc., New York, 1993.
[15] D.A Lange, S. Penuela, L.R. Denny, J. Mudge, V.C.
Concidido, H.J. Orf, et al., A plant isolation protocol
suitable for polymerase chain reaction based
marker-assisted breeding, Crop Science 138 (1998)
217-220.
[16] A. Dresher, A. Grane, PCR-genotyping of barley seedling
using DNA samples from tissue prints, Plant Breeding 121
(2002) 228-231.
Journal of Agricultural Science and Technology B 2 (2012) 368-377
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Evaluating the Impact of Adoption of Improved Maize

Varieties on Yield in Uganda: A Propensity Score
Matching Approach

Margaret Nabasirye1, Barnabas Kiiza2 and George Omiat2

1. Department of Agricultural Production, Makerere University, P.O. Box 7062, Kampala, Uganda
2. Department of Agribusiness and Natural Resource Economics, Makerere University, P.O. Box 7062, Kampala, Uganda

Received: January 19, 2012 / Published: March 20, 2012.

Abstract: The main objective of this study was to evaluate the impact of adopting improved maize varieties on crop yield in Uganda
using propensity score matching (PSM) algorithms to control for hidden selection bias. The study employed cross-sectional household
data collected in 2005/2006 by the Uganda Bureau of Statistics (UBOS) across the country. Data were divided into two parts; the full
sample and smallholder farmer sub-sample (those that cultivated less than 5 acres or 2 Hectares of maize in 2004 and 2005). Then
estimation was made of the difference in yields between the adopters and non adopters, the average treatment effect on the treated (ATT)
for the full sample and smallholder farmer sub-sample. Matching methods were used to control for hidden selection bias and the
standardized bias measure was used to check for the quality of matching. The results indicate that adoption of improved maize seed had
a robust positive and significant effect on yields obtained by farmers. The results were consistent for both the full and smallholder
farmer sub-samples. Sensitivity analysis using Rosenbaum bounds indicated that the ATT results were fairly robust to hidden selection
bias. That is, the results were not sensitive to unobserved selection bias. Therefore it is confidently asserted that the estimated average
difference in maize yields between the adopters and non-adopters was due to the effect of planting improved maize seed.

Key words: Adoption, impact evaluation, improved maize seed, propensity score matching, Uganda.

1. Introduction much between 1990 and 2007. The potential maximum

yield is about 6.0 metric tons/hectare [3]. Maize
This study used the maize commodity to establish
provides over 40% of the calories consumed by
the importance of planting improved seed in relation to
households in rural and urban areas of the country.
food security and income enhancement among
Many households grow maize as a food security crop
smallholder farm households. Maize is a major food
[4]. To increase yield of maize in the country national
crop in the country, and is grown in almost all districts,
breeding programs, together with CIMMYT and other
and is easy to manage, relatively resistant to water
international organizations, have introduced hybrid and
stress, and adaptable to different soil types [1].
composite maize varieties which have been
According to RATES [2], maize provides a living for
disseminated and adopted widely. There are many
about two million households and accounts for about
studies on the adoption of improved maize and the
95% of smallholder farm production. Secondary data
impact on yields in Africa [5-8]. The work reported
show that the national average yield for maize has been
here differs from the types of previous studies
around 1.50 metric tons/hectare and has not increased
mentioned above in that it employs PSM (propensity
Corresponding author: Margaret Nabasirye, Ph.D., research
score matching) methods to effectively control for
fields: biometrics, applied statistics, research methods. E-mail: hidden selection bias which has not been done by most
mnabasirye@agric.mak.ac.ug.
 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
of the previous studies. The objective of the present
study was to determine the impact of adoption of
improved maize on yields in Uganda. Using PSM,
adopters were matched with non-adopters on a number
of characteristics to ensure that the only difference
between the two cohorts was the fact that one cohort
had adopters and the other had non-adopters of
improved maize seed. Improved maize seed considered
here is of hybrid and composite varieties.
2. Materials and Methods
2.1 Analytical Framework
There are many studies that evaluate the impact of
adopting improved agricultural technologies which
fail to properly control for potential differences
between the technology participants/adopters and
non-participants. This makes it difficult to draw
definitive conclusions about the magnitude of the
impacts [9-10]. A simple analysis of differences in the
average yield for maize between the adopters and
non-adopters more often than not produces
statistically misleading results. Farmers choose to
adopt or not adopt a given technology, depending on
their expectations, objectives, and observable and
unobservable characteristics. This is referred to as
self-selection in impact literature. For instance,
farmers who adopt improved maize varieties may be
the ones who are wealthier, more endowed, more
educated, or more experienced than the non-adopters
and these characteristics may not be obvious to the
researcher. As a result of this, a comparison of
adopters with non-adopters tends to overestimate the
impact of the technology on yields obtained. In
addition, some farmers may disproportionately select
into the adoption program due to their higher
propensity to participate in technology adoption, and
this too is unobservable to the researcher. Therefore,
in absence of random selection of farmers in the
adoption of improved maize seed, simple
comparisons of average yields obtained between
adopters and non-adopters are likely to give upwardly
369
biased estimates of the impact of adoption [11-13]. To
control for possible hidden selection bias,
econometric methods usually employ Heckman
two-step estimation. Heckman models have the
advantage that they are able to model differences in
both observable and unobservable characteristics
between the treated and control groups by inclusion of
the inverse Mills ratio as an extra explanatory variable
in the outcome model [12]. However, Heckman
selection estimators are based on the strong
assumption that unobserved variables are normally
distributed and their robustness has been questioned
in a growing literature using both simulated and actual
data. Selection bias can also be controlled using the
IV (instrumental variables) method which can
produce consistent estimation of causal effects in the
presence of hidden bias. The IV approach eliminates
the selection bias problem, but the method is difficult
to use because it is difficult to find instruments that
affect the probability to participate but do not affect
outcome. Furthermore, the IV method typically
reduces the precision of causal estimates and
introduces new uncertainty from its reliance on
additional untestable assumptions [14].
Another drawback of the IV method is the
requirement of at least one variable in the selection
model to serve as an instrument in specifying the
outcome model. It is not easy to find good instruments
in empirical studies. Jalan and Ravallion [15] and
Mendola [10] also indicated that this method imposes
an assumption of a linear functional form which may
not always hold. That is, the coefficients on the control
variables are similar for adopters and non-adopters.
Matching methods are also used to address the
self-selection bias. Several matching methods have
been developed. A commonly used matching method is
PSM, which matches the adopters and non-adopters
based on the observable characteristics that affect
participation and the outcome being measured. PSM,
originally used by Rosenbaum and Rubin [16], has
been employed widely [10-12, 17-22]. Moreover, PSM
370 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
does not require exclusion restrictions or a given
specification of the functional form of the selection
model to construct the counterfactual case or to reduce
self-selection bias.
2.2 Propensity Score Matching Method
Recent statistical literature on causal estimation has
made the definition of an average causal effect much
more precise with the knowledge that causal effects are
typically heterogeneous in a population [14]. The
causal effect of the treatment varies from individual to
individual due to the interpersonal variation in
observable variables. DiPrete and Gangl [14] further
stated that there is more than one way of computing an
average (of which the average treatment effect for the
treated, the ATT, is one). In empirical work, it is only
the average causal effects that can be identified for
analysis and the researcher is faced with the problem of
generating an appropriate estimator for the causal
effect of interest. Therefore, standard regression
models are not satisfactory because they do not
rigorously define any specific average treatment effect.
Rosenbaum and Rubin [16-17] showed that the
method of PSM overcomes this problem by matching
only one variable, the propensity score, which is
estimated as a probability via a probit or logit selection
model. They indicated that matching with reference to
the propensity score is equivalent to matching
sequentially on each of the individual explanatory
variables that determine the outcome variable. The
influence of confounding variables is reduced by the
method of covariate balancing, matching the
potentially confounding explanatory variables of the
individuals who experienced the treatment with
individuals who did not experience the treatment (the
control group). In this way, PSM eliminates any
relationship between the covariates and assignment to
the treatment, and eliminates the possibility of bias
from these variables.
Following the standard model requires first to
examine the issue of constructing the counterfactual
using PSM methods. There are adopters and
non-adopters of improved maize varieties. The
counterfactual is what would have happened to those
farmers who, in fact, adopted improved maize varieties,
had they not adopted the technology. Thus, of specific
interest is to know what would have happened to the
adopters, had they not adopted the technology. The key
assumption is that individuals selected into treatment
and non-treatment (control) groups have potential
outcomes in both states, the one in which they are
observed and the one in which they are not observed
[23]. Let D = 1 denote the state when the ith farmer
adopts improved maize seed, and D = 0 denote the state
when he does not adopt improved maize seed. Let Yi
denote the actual observed outcome of individual
farmer i. Then Y1i is the outcome of ith farmer when he
adopts the technology and Y0i is the outcome if the
farmer does not adopt the technology. The outcomes
for the ith farmer can be defined as:
Y1 = 1X + u1 and (1)
Y0 = 0X + u0
where X is a vector of observed covariates, and u1 and
u0 are unobserved random error terms with the
assumption that u1 u0 and the effect of choosing
adoption versus no adoption is given as (Y1 Y0).
Following Heckman et al. [24], the outcome equation
for observed Y, conditional on treatment participation,
may be written as a switching regression:
Y = DY1 + (1 D)Y0 (2)
= 0 + D(1 - 0 + u1 - u0) + u0
= 0 + D(1 - 0) + [D( u1 - u0) + u0].
In reduced form, Eq. (2) can be rewritten as a
standard regression model:
Y = 0 + 1D + u (3)
However, the treatment effect, (Y1 Y0) cannot be
estimated using the model in Eq. (3) due to the strong
potential of selection bias, thus turning to PSM to
determine the average treatment effect on the treated
farmers (the ATT). That is, the causal effect of
adoption of improved maize seed on yields. For a given
adopter, consider both the observed mean yield under
 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
the condition of adoption of the technology as E (Y1 |D
= 1) and the unobserved (hypothetical) mean yield that
the adopter would have realized had they not adopted
improved maize seed as E (Y0 |D = 1). Similarly, for a
given non-adopter, consider the observed mean yield
under the condition of non-adoption of the technology
as E (Y0 |D = 0) and the unobserved (hypothetical)
mean yield that the non-adopter would have realized
had they indeed adopted improved maize seed as E (Y1
|D = 0). Here, E is the expectation operator.
Rosenbaum and Rubin [16] and Caleindo and Kopeinig
[19] showed that the parameter of interest is the ATT,
which is written as:
ATT = E(Y1 - Y0 |D = 1)
= E(Y1 |D = 1) - E(Y0 |D = 1) (4)
The central interest of impact evaluation is not in E
(Y0 |D = 0), but in E (Y0 |D = 1). Therefore, PSM uses
balancing scores to extract the observed mean yield of
those non-adopters who are most similar in observed
characteristics to the adopters. That is, it uses E (Y0|D =
0) to estimate the counterfactual, E (Y0 |D = 1).
Estimation for the true parameter requires that
E (Y0 |D = 1) E ( Y0 |D = 0) = 0 (5)
This ensures that the ATT is free of self-selection
bias. To fulfill the condition in Eq. (5), PSM must
satisfy two assumptions: conditional independence and
common support. The conditional independence
assumption (also known as exogeneity) requires that
the value of the outcome variable is independent of the
treatment state, given the values of some observable
variables [16, 17]. That is, selection into the treatment
group is solely based on observable characteristics. The
orthogonality condition expressed in Eq. (6) says that
the values of the outcome variables Y1i and Y0i are
independent of the treatment state (D), given the values
of the observable variables in X.
Y1i, Y0i D | X (6)
Any systematic effect of treatment on the outcome
variable (in this case maize yield) can be entirely
explained in terms of these observables. Thus, the
differences in unobservable characteristics between the
371
treated and untreated groups conditional on X are
assumed to be random.
The common support assumption is needed since
matching on every covariate is difficult when there are
many covariates. PSM solves this problem by
estimating the propensity score which is given as P(x)
= Pr (D = 1| X). This is the conditional probability that
the ith farmer will adopt improved maize seed,
conditional on the observed characteristics in vector X.
To ensure that every individual has a positive
probability of being both an adopter and non-adopter,
thus ruling out perfect predictability, the common
support (overlap) condition implies that
0 < Pr (D = 1|X) < 1 (7)
These two conditions give us the strong ignorability
of treatment assumption that allows us to use the PSM
estimator. They imply that
E(Y0 |D = 1, P(x)) = E( Y0 |D = 0, P(x)) (8)
Then, according to Rosenbaum & Rubin [16], the
ATT can be written
ATT = E(Y1 - Y0 |D = 1) (9)
= E [E(Y1 - Y0 |D = 1, P(x))]
= E [E(Y1 |D = 1, P(x)) - E(Y0 |D = 1, P(x)) | D = 1]
= E [E(Y1 |D = 1, P(x)) - E(Y0 |D = 0, P(x)) | D = 1]
where the outer expectations are over the distribution
of Pr (D = 1|X).
2.2.1 Quality of Matching (Covariate Balancing)
According to Rosenbaum and Rubin [16], formation
of matched pairs of observationally similar adopter and
non-adopter cases eliminates the confounding effects
of observable variables. Matching is intended to restrict
the control sample (non-adopters) in order to increase
the similarity of the sub-sample of control cases that
are directly compared with the treated cases (adopters)
in order to estimate the unbiased impact of the adoption
of improved maize seed. Therefore, the quality of
covariate balancing (equality of means on the p-scores
and equality of means on all covariates) between
adopters and non-adopters is tested using the
standardized bias between treatment and non-treatment
samples, as suggested by Rosenbaum and Rubin [17,
372 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
18] and Rubin [19]. For each variable X and propensity
score, the standardized bias SB(X) is computed before
and after matching as:
SB ( X )
100 X T X C (10)
2
T 2
C / 2
where X T and X C are the sample means for the
treatment and control groups, and 2T and 2C are the
corresponding variances [20]. The post-matching bias
reduction (BR) can be computed as:
SB POST (11)
BR 100 1
SB PRE
The BR ratio shows the percentage reduction in
standardized bias due to matching. That is, the
standardized bias measures the difference between
covariate means of the treatment and control groups in
terms of the number of standard deviations from zero
[25]. Other covariate balancing indicators used are the
likelihood ratio test of the joint significance of all
covariates and the pseudo-R2 from a logit of treatment
status on covariates before matching and after
matching on the matched sample. After matching, there
should be no systematic differences in the distribution
of covariates between the two groups. This implies that
the pseudo-R2 should be fairly low and the joint
significance of all covariates can be rejected.
2.2.2 Robustness Tests (Rosenbaum Bounds)
PSM estimators are not consistent estimators for
treatment effects if the assignment to treatment is
nonrandom, i.e., if there are unobserved covariates that
affect the assignment to treatment and are also related
to the outcomes [14, 26]. An unobserved covariate that
affects assignment to treatment but that does not affect
the outcome beyond the covariates already controlled
for, does not challenge the robustness of the estimation.
In order to estimate the extent to which such selection
on unobservable covariates may bias the inferences;
this study employed the Rosenbaum bounds sensitivity
analysis to determine how strongly a non-measured
variable must influence the selection process so that it
could undermine the implications of the matching
analysis. If there is a certain unobserved variable of
concern that affects the selection process then the
probability of treatment is
P(X) = Pr (D = 1|X, u) = F (X + u) (12)
where X is a vector of all observed covariates and u
represents the unobserved variable affecting
assignment to treatment; is the effect of u on the
treatment probability. If the estimator is free of hidden
bias, is equal to zero and the participation probability
is solely determined by X. However, if there is hidden
bias, two individuals with the same observed covariates
can have different chances of adopting improved maize
seed. For example, consider a matched pair of
individuals i and j, and logistic distribution F. The odds
that an individual receives a treatment is given by P/(1
P). This means that for a matched pair of individuals
(i and j) the odds ratio is shown in Eq. (13), where the
and X variables are vectors.
Pi 1 Pi exp X i u i (13)
P j 1 P j exp X j u j
Since the vector Xi = Xj , the odds ratio reduces to
exp [(ui uj )]. The bounds on the odds ratio that
either of the two individuals receives the treatment is
1 P 1 Pi (14)
i e
e P j 1 P j
Here e is denoted as , as is the standard case. If =
1, there will be no hidden bias as either = 1 or ui = uj.
That is, if = 1, individuals with the same observable
characteristics have the same probabilities and odds of
assignment to treatment. If = 2, then one individual
with the same observable characteristics as another is
twice as likely to be selected into the treated group. The
bounds on the odds ratio start to vary as takes on
values greater than one. The Rosenbaum bounds at
each value of is the point at which hidden bias causes
us to question the findings. This cut-off point should be
large enough to render the estimates robust against the
presence of unobservable selection bias. Comparison
of the Rosenbaum bounds on treatment effects at
different levels of can assess the strength that
unmeasured covariates must have in order for the
estimated treatment effects from propensity score
 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
matching to have arisen purely through nonrandom
assignment [14].
2.3 The Data and Study Area
The data for this study were obtained from the
2005/2006 national household survey conducted by the
UBOS. About 7,400 households were interviewed in
the survey country-wide. About 1,000 households were
reported in the survey to have grown maize, both
traditional and improved varieties. This sample size
was sufficiently large and diverse to represent the
households of interest in this study. The survey had an
agriculture module and a socio-economic module. The
details of each of the modules are summarized below.
The agricultural module covered the household crop
farming enterprise particulars with emphasis on
investments on land, crop areas, labor and non-labor
inputs for both the first and second seasons, crop
harvests, livestock ownership, expenditure on livestock
and agricultural extension services and technologies.
The socio-economic module covered household
characteristics including education and literacy, health
status and behavior of household members, activity
status of household members, wage employment,
enterprise activities, transfers and household incomes,
housing conditions, assets, loans, household
expenditure, welfare indicators and household shocks.
2.3.1 Treatment and Outcome Variables
The adopters of improved maize seed are farmers
who obtained and planted improved maize in season
two of 2004 and season one of 2005. The non-adopters
are those who use recycled or farm-saved seed from the
previous harvests in seasons two of 2004 and season
one of 2005 and did not plant improved maize seed.
The binary treatment indicator is D = 1 for adopters and
D = 0 for non-adopters. Only data for season one of
2005 were used for this study. Maize yields are
measured as the ratio of output of maize to the area
cultivated in the first season of 2005. That is, kilograms
of maize obtained per acre. The covariates are variables
that can predict both adoption of improved maize seed
373
and the outcome variable, maize yield. These include,
inter alia; age of the farmer, sex, level of education,
type of occupation, type of dwellings owned, acreage
of land owned, wealth, distance of farm from district
town or main market, non-farm income, work
experience, membership in a farmer organization,
participation in government technology awareness
campaigns, access to microfinance loans, if farmer had
training on use of improved seed before adopting, etc..
Regional dummy variables are included to capture
location specific characteristics. Reported in Table 1
below are descriptive statistics for the variables
separated according to the treatment and non-treatment
cases for seed adoption.
3. Results and Discussion
3.1 Testing Quality of Matching
Testing for the quality of matching the adopters and
non-adopters was made through covariate balancing
tests. The quality of the radius and Epanechnikov
kernel matching results are given in Table 2. After
matching there is a reduction in the following: the
standardized bias; the pseudo-R-square; the likelihood
ratio Chi-square; and a reduction in the statistical
significance the likelihood ratio Chi-square as shown
by the p-value increase. This indicates that after
matching there are no systematic differences in the
distribution of covariates in the vector X between the
treated and the control groups. That is, the hypothesis
that both groups have the same distribution of the
covariates in X after matching cannot be rejected.
Propensity score distribution and common support for
propensity score estimation figures are all available
upon request from the authors.
In this study the first matching statistic in testing the
quality of covariate balancing (equality of means on all
covariates) between treatment and control cases is the
standardized bias (SB) reduction test. It is the reduction
in bias after matching. Table 2 shows that the mean
bias reduction ranges between 39.29 percent and 46.59
percent after matching. This shows a good level of
374 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach

Table 1 Descriptive statistics: covariates used in the study.

Treatment cases Non-treatment cases
Variable (improved maize seed) (traditional maize seed)
N Mean S.D N Mean S.D
Membership in farmer association = 1; Otherwise = 0 541 0.0684 0.2527 2,951 0.0502 0.2183
Age of the household head (hhh) 541 43.7124 14.23282 2951 44.1306 15.69873
Years of formal education of the household head. 541 6.1823 3.53815 2951 4.9498 3.50285
Sex of the household head. Male = 1; Female = 0 541 0.8145 0.38909 2951 0.7321 0.44293
Rural location = 1; Otherwise = 0 541 0.86 0.348 2951 0.90 0.299
Accessed microfinance loans = 1; Otherwise = 0 541 0.1905 0.39385 2951 0.1173 0.32195
Number of rooms in the house 541 7.4030 3.92762 2951 6.7014 3.91653
Iron sheet roof = 1; Grass thatched = 0 541 0.6654 0.47228 2951 0.5613 0.49632
Brick walls house = 1; Mud walls = 0 541 0.3494 0.47721 2951 0.3182 0.46586
Cement floor = 1; Earth Floor = 0 541 0.2292 0.42071 2951 0.1659 0.37203
Owns mobile phone = 1; Otherwise = 0 541 0.1885 0.39150 2951 0.1012 0.30167
Owns radio = 1; Otherwise = 0 541 0.6969 0.46004 2951 0.6523 0.47631
Owns bicycle = 1; Otherwise = 0 541 0.5231 0.49993 2951 0.5183 0.49975
Household size 541 6.5786 3.34163 2951 6.0091 3.08722
Obtained extension services = 1; Otherwise = 0 541 0.1553 0.36250 2951 0.0613 0.23998
Trained in agronomic practices = 1; Otherwise = 0 541 0.1238 0.32971 2951 0.0830 0.27596
Land owned (Acres) 541 9.2534 42.26291 2951 7.7514 78.19781
Value of livestock owned (Shillings) 541 160,050 703,977 2951 68,422 463,888
Off farm income (Shillings) 541 534,694 1,901,571 2951 283,326 1,113,403
Farm income (Shillings) 541 785,219 1,555,639 2951 310,374 803,105
Price of maize obtained (Shillings / Kg) 541 221 294 2951 276 383
Used manure = 1; Otherwise = 0 541 0.0598 0.23736 2951 0.0189 0.13629
Used fertilizer = 1; Otherwise = 0 541 0.0841 0.27782 2951 0.0058 0.07627
Used pesticides = 1; Otherwise = 0 541 0.0579 0.23386 2951 0.0193 0.13752
Exposed to govt awareness campaigns on improved 541 0.1537 0.36100 2951 0.0780 0.26816
seed = 1; otherwise = 0
Regional Dummies
Central = 1; Otherwise = 0 541 0.17 0.378 2,951 0.26 0.439
Eastern = 1; Otherwise = 0 541 0.62 0.485 2,951 0.31 0.463
Northern = 1; Otherwise = 0 541 0.12 0.325 2,951 0.21 0.405
Western = 1; Otherwise = 0 541 0.09 0.279 2,951 0.22 0.414
Source: UNHS data; Exchange rate US$1.00 = Ugandan Shillings (Shs) 1,800.

Table 2 Covariate balancing tests for the effect of improved maize on yields (kg/acre).
Outcome variable Bias reduction Pseudo R-squared LR chi-square P-value of LR
Maize yield
Full sample
0.274 198.55 0.000
Kernel matching 46.59%
0.085 32.65 0.140
0.274 198.55 0.000
Radius matching
39.29% 0.029 20.81 0.703
Smallholder sub-sample
0.272 175.53 0.000
Kernel matching
42.98% 0.036 23.23 0.564
39.78% 0.272 175.53 0.000
Radius matching
0.039 24.89 0.468
Numbers in bold italics are the before-matching statistics. The numbers in regular font are the after-matching statistics.
 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
covariate balancing. There is no formal criterion in the
literature on the magnitude of the standardized bias
reduction. However, Rosenbaum and Rubin [17]
suggested that a value of 20% bias reduction after
matching is large enough. The second matching
statistic employed is the pseudo-R-square from the
logit estimation of the conditional probabilities. It
indicates the degree to which regressors in the vector X
predict the treatment probability. After matching, the
regressors in the vector X should have no or very low
explanatory power for selection into treatment if the
treatment and matched control samples have balanced
characteristics. The results show that this is the case.
The pseudo-R-square statistic drops after matching. It
drops to between 0.029 and 0.085 after matching. The
third measure of covariate balancing is given by the
p-values of the likelihood ratio Chi-square (LR) before
and after matching. Results indicate that after matching
there are no systematic differences in the distribution of
covariates in the vector X between the adopters and
non-adopters. That is, the hypothesis that both groups
have the same distribution of the covariates in X after
matching cannot be rejected as shown by the increase
in the P-values of the likelihood ratio Chi-square. This
suggests that there is no systematic difference in the
distribution of covariates between adopters and
non-adopters after matching.
3.2 Average Treatment Effect on the Treated (ATT)
In this study PSM was used to check the existence
and robustness of the causal interpretation given to the
association found between adoption of improved maize
varieties and higher yields obtained by farmers who
have adopted improved seed. This causal impact of
adoption of improved maize seed on farm yields is
done by selection of a very large number of observable
factors. The robustness of the causal impact is done
after controlling for selection on unobservable or
confounding variables or characteristics (hidden
selection bias). That implies the authors can
confidently attach causal interpretation to the impact of
375
adoption of improved maize seed on maize yields
obtained by farmers.
The ATT is the statistic of interest in this study. That
is, the difference in yields between the adopters and
non-adopters after controlling for hidden selection bias.
Results in Table 3 indicate the ATTs obtained using
radius and Epanechnikov Kernel matching algorithms.
These results indicate that adoption of improved maize
has a robust positive and significant effect on maize
yields. Becerril and Abdulai [27] obtained similar
results. This effect of adoption of improved maize seed
is consistent for both the entire sample and smallholder
sub-sample. From the full sample, a sub-sample of
farmers was extracted to denote smallholders who, on
average, cultivated less than 5 acres of maize in season
one of 2005. Results in Table 3 show that, on average,
the increase in maize yields after adoption of improved
seed (ATT) is about 371 kg/acre in the full sample in
2005 using the Epanechnikov kernel matching
algorithm. It is about 359 kg/acre using the radius
matching algorithm. In the smallholder farmer
sub-sample, on average, the increase in maize yields
after adoption of improved seed is about 221 kg/acre
using Epanechnikov kernel matching. It is about 230
kg/acre using radius matching algorithm.
3.3 Robustness Tests
Although Rosenbaums procedure addresses
situations of both positive and negative self-selection
into treatment status, the case of positive self-selection
was the one of primary interest in this study. Therefore,
subsequent investigations were restricted to bounds for
positive self-selection. Sensitivity analysis using
Rosenbaums bounds in the full sample shows the
critical values of range between 1.35 and 1.45 at the
5% and 10% levels of significance, respectively using
Epanechnikov kernel matching. The range is 1.25 and
1.35 at the 5% and 10% levels of significance,
respectively using radius matching. The results for the
smallholder sub-sample indicate that the range of is
between 1.30 and 1.45 at the 5% and 10% levels of
376 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
Table 3 ATT results: Effect of improved maize on yields (kg/acre).
Full sample
Matching algorithm ATT t -value Critical level of for hidden bias
1.35
Kernel matching 370.53 2.61***
1.45
1.25
Radius matching 358.96 2.55**
1.35
Smallholder farmer sub-sample
ATT t-value Critical level of for hidden bias
1.30
Kernel matching 220.69 3.31***
1.45
1.35
Radius matching 229.72 3.31***
1.50
Figures in Italics indicate at 5% level. The others show at 10% level of significance.
***, **, * indicates significance at 1%, 5%, and 10%, respectively.
significance, respectively using Epanechnikov kernel
matching. The range is 1.35 and 1.50 at the 5% and
10% levels of significance, respectively using radius
matching. The present study considered the
Epanechnikov kernel matching results in all cases since
it is more conservative method.
For the smallholder farmers, these results that show
= 1.45 at 10% level of significance, imply that doubts
over the statistical significance of the estimated ATT
becoming insignificant would emerge only if a
confounding factor causes the odds ratio of assignment to
the adopter cohort to differ between the two cohorts by a
factor of 1.45 as indicated by DiPrete and Gangl [14].
Even if the odds ratio of assignment to treatment was
allowed to differ between adopters and non-adopters
by a factor of about 1.40, the effect would still not be
sensitive to hidden selection bias due to unobserved or
confounding covariates. Therefore these results are
fairly robust to hidden selection bias. The authors can
confidently assert that adoption of improved maize
seed has a fairly robust positive and significant impact
on maize yields obtained by the farmers.
4. Conclusions
This study employed propensity score-matching
methods to determine the impact of adoption of
improved maize varieties on yields obtained by farmers
by controlling hidden selection bias. In both the full
and smallholder maize farmer sub-samples, adoption
Number of matched pairs
138
138
Number of matched pairs
119
119
of improved maize varieties had a positive and
significant effect on yields obtained by the farmers.
Thus adoption of improved maize technology has
positive implications for food security and poverty
alleviation in Uganda. Yield-augmenting agricultural
technologies are among the major ways by which food
security and poverty alleviation can be achieved,
especially for smallholder farm households. These are
farmers who use very little agro-chemicals such as
fertilizers and pesticides or irrigation and thus obtain
very low yields when planting traditional maize
varieties.
References
[1] K. Deininger, J. Okidi, Rural household incomes,
productivity and non-farm enterprises in Ugandas
recovery: The role of farms, firms and government, in:
Reinikka and Collier (Eds.), The World Bank, Washington,
D.C., 2001.
[2] RATES, Maize Market Assessment and Baseline Study
for Uganda, The RATES Center, Nairobi, Kenya, 2003.
[3] Uganda Bureau of Statistics (UBOS), Uganda National
Household Survey 2005/2006, 2007.
[4] B.P. Mungyereza, Impact of Agricultural Market and Price
Liberalization on Food Security in Uganda: A Policy
Analysis, Coopibo-Uganda, SNV-Uganda,
Oxfam-Uganda, 1998.
[5] M. Morris, R. Tripp, A. Dankyi, Adoption and impacts of
improved maize production technology: A case study of
the Ghana grains development project, Economics
Program Paper 99-01, Mexico, D.F., CIMMYT, 1999.
[6] G. Mwabu, W. Mwangi, H. Nyangito, Does adoption of
improved maize varieties reduce poverty?, Evidence from
 Evaluating the Impact of Adoption of Improved Maize Varieties on Yield
in Uganda: A Propensity Score Matching Approach
Kenya, in: The International Association of Agricultural
Economists Conference, Gold Coast, Australia, August,
2006.
[7] A.S. Oyekale, E. Idjesa, Adoption of improved maize seed
and production efficiency in Rivers State, Nigeria,
Academic Journal of Plant Sciences 2 (2009) 44-50.
[8] J. Mugisha, G. Diiro, Explaining the adoption of improved
maize varieties and its effects on yields among
smallholder maize farmers in eastern and central Uganda,
Middle-East Journal of Scientific Research 1 (2010) 6-13.
[9] S. Rahman, Impact of technological change on income
distribution and poverty in Bangladesh agriculture: An
empirical analysis, Journal of International Development
11 (1999) 935-955.
[10] M. Mendola, Agricultural technology adoption and
poverty reduction: A propensity-score matching analysis
for rural Bangladesh, Food Policy 32 (2007) 372-393.
[11] M. Kassie, B. Shiferaw, G. Muricho, Adoption and impact
of improved groundnut varieties on rural poverty: Evidence
from rural Uganda, Environment for Development
Discussion Paper Series EfD DP, 2010, pp. 10-11.
[12] B. Kiiza, G. Pederson, S. Lwasa, The role of market
information in adoption of agricultural seed technology in
rural Uganda, International Journal of ICT Research and
Development in Africa 2 (1) (2010) 29-46.
[13] P. Holland, Statistics and causal inference, Journal of the
American Statistical Association 81 (396) (2007) 945-960.
[14] T. DiPrete, M. Gangl, Assessing bias in the estimation of
causal effects: Rosenbaum bounds on matching estimators
and instrumental variables estimation with imperfect
instruments, Sociological Methodology 34 (2004)
271-310.
[15] J. Jalan, M. Ravallion, Does piped water reduce diarrhea
for children in rural India?, Journal of Econometrics 112
(2003) 153-173.
[16] P.R. Rosenbaum, D.B. Rubin, The central role of the
propensity score in observational studies for causal effects,
Biometrika 70 (1983) 41-55.
377
[17] P.R. Rosenbaum, D.B. Rubin, Constructing a control
group using multivariate matched sampling methods that
incorporate the propensity score, The American
Statistician 39 (1985) 33-38.
[18] D.B. Rubin, Practical implications of the modes of
statistical inference for causal effects and the critical role
of the assignment mechanism, Biometrics 47 (1991)
1213-1234.
[19] R.H. Dehejia, S. Wahba, Propensity score-matching
methods for non-experimental causal studies, Review of
Economics and Statistics 84 (2002) 151-161.
[20] M. Caliendo, S. Kopeinig, Some practical guidance for the
implementation of propensity score matching, Journal of
Economic Surveys 22 (2008) 31-72.
[21] J. Smith, P. Todd, Does matching overcome LaLondes
critique of nonexperimental estimators?, Journal of
Econometrics 125 (2005) 305-353.
[22] A. Akhter, A. Abdulai, The adoption of genetically
modified cotton and poverty reduction in Pakistan, Journal
of Agricultural Economics 61 (2010) 175-192.
[23] B. Winship, S. Morgan, The estimation of causal effects
from observational data, Annual Review of Sociology 25
(1999) 659-706.
[24] J. Heckman, S. Urzua, E. Vytlacil, Understanding
instrumental variables in models with essential
heterogeneity, Review of Economics and Statistics 88
(2006) 389-432.
[25] D. Harding, Counterfactual models of neighborhood
effects: The effect of neighborhood poverty on dropping
out and teenage pregnancy, American Journal of
Sociology 109 (2003) 676-719.
[26] S. Becker, A. Ichino, Estimation of average treatment
effects based on propensity score, The Stata Journal 2 (4)
(2007) 358-377.
[27] J. Becerril, A. Abdulai, The impact of improved maize
varieties on poverty in Mexico: A propensity
score-matching approach, World Development 38 (2009)
1024-1035.
Journal of Agricultural Science and Technology B 2 (2012) 378-384
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

Influence of Frozen Storage Time on Unsaturated Fatty

Acids and Proximate Composition of Caspian Sea White
Fish (Rutilus frisii kutum) Fillets

Sahar Jalili
Department of Fisheries, Abadan Branch ,Islamic Azad University, Abadan, Iran

Received: August 22, 2011 / Published: March 20, 2012.

Abstract: The main aim of this research was to identify fatty acids composition of Caspian Sea of white fish Rutilus frisii kutum tissue
and their changes during 12 months storage (-18 C ). The results showed, UFA (Unsaturated Fatty Acid) and SFA (Saturated Fatty
Acid) were 74/09 and 21/63%, respectively in fresh tissue. So that, DHA (C22:6) oleic acid (C18:1c) had high amounts (15/07, 20/57)
UFA and palmitic acid (C16:0) was the most (13/09%) SFA. The effects of freezing on fish tissue showed that UFA and SFA contents
have reached to 58/79 and 22/17%, respectively at the end of cold storage. Also -3 and -6 series of fatty acids was 24/22 and 15/56%
in fresh tissue, but their contents decreased to 8/68 and 5/11% at the end of period. Among, the fatty acids C22:6, C18:1c and C16:0 had
the most changes. The changes of fatty acids were significantly at 95% level expected for C18:0.

Key words: Caspian Sea, Rutilus frisii kutum, frozen storage, -3, -6.

1. Introduction extent on the method of storage. Storage time and

temperature are the major factors affecting the rate of
In last decades, PUFAs (polyunsaturated Fatty
loss of quality and shelf life of fish. In spite of some
Acids) of 3 family have been recognized to be
disadvantages associated with frozen storage, freezing
essential components of human diet. These acids,
is accepted as effective way of preserving fish [4-5].
particularly EPA (Eicosopentaenoic ) 20:5 3 and DHA
Mean while, unsaturated fatty acids are more
22:6 6 , appeared to play key role in ontogenesis,
susceptible oxidation than their saturated analogous,
especially neural development functioning of
and PUFAs content in aquatic species was
cardiovascular system and immune system. Regular
demonstrated to decrease during storage [6]. The study
consumption of food with appropriate content of EPA
therefore aimed at determining the acceptable storage
and DHA provides prevention and treatment of
life of frozen fillets and the freezing effects on
depression cardiovascular and some other disease [1-2].
chemical composition fatty acid profiles of the fish
Researchers have found that, fresh water fish contain
commonly consumed, in Iran.
lower proportions of n-3 PUFA than do marine fish.
Therefore, the ratio of total n-3, n-6 fatty acids is much 2. Materials and Methods
higher for marine fish than for fresh water fish,
2.1 Sample Collection
verifying from 5 to 10 or more [3]. However,
availability of these vital nutrients depends to a large Kutum fish (Rutilus frisii kutum) average weight
(1,000 100 g) used in study, were caught in March 15
Corresponding author: Sahar Jalili, Ph.D., assistant of 2010 from Caspian Sea. After being caught, they
professor, research field: fishery science. E-mail: were transferred to laboratories, filleted and then
Sahar.jalili2005@gmail.com.
 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets
frozen at -30 C. Frozen fillets, dispatched packed into
a box with ice to the Faculty of Veterinary Medicine,
University of Tehran for analyses. Fish (Rutilus frisii
kutum) were carried out 1, 3, 6, 9 and 12 months from
beginning storage.
2.2 Proximate Composition Analysis
Water content was determined by the sample drying
technique at a temperature of 105 C; crude protein
content with the use by the Kjeldahl method with 6.25
muliplier, fat content was determined with Blight and
Dyer method and ash content by sample mineralization
at a temperature of 550-600 C [3].
2.3 Determination of Fatty Acid Profile
Total lipid was extracted by chloroform-methanol
and estimated (1:1, v/v) and estimated gravimetrically
[7]. The fatty acids in the total lipids were esterified into
methyl esters by saponification with 0/5 N methanolic
NaOH and trans-esterified with BF3(Tri Fluoride Bore)
(w/v) methanol [8]. The FAMEs were analyses on
HP5890 Series II GC (Gas Chromatograph) equipped
with gas chromatograph with a FID (Flame Ionizing
Detector) and a 30 m capillary column with an internal
diameter of 0.25 mm, liquid phase Supelco wax 10, film
thickness 0.25 m. Separation conditions were:
nitrogen carrier gas, flow rate of 1 L/min.
Temperature of the detector was 260 C, of the injector
250 C; and of the column 155 C. The individual fatty
acids were identified by comparing their retention times
with the standards of Supelco.
2.4 Sensory Assessment
Fillets with skin to be used for sensory assessment
were evaporated to an internal temperature of 72 C.
Immediately after the thermal treatment, 5 cm of the
fillet was cut out beginning with the section touching
the head next, 6 samples of 2.0 cm wide were cut out
from the central section. A sample evaluated by each
assessor originated from the same part of the fillet.
Procedure was designed to minimize potential
differences likely to result from the evaluation of
379
different parts of the fillet. The assessment was carried
out by 6 person panel, and all assessors evaluated all
samples examined (fresh, frozen kutum fish). The
sensory evaluation was carried out in three separate
sessions. The assessors performing the sensory
assessment met the minimum sensory competence
required in the current calibrating tests; they were
selected, trained, tested, and monitored according to
ISO guidelines [9].
The assessors were experienced in sensory
evaluation and were acquainted with the accepted
system of sensory assessment. Color, odor, consistency,
juiciness and flavor were determined on a 5-point scale
using half-scores according to the modified method for
the evaluation of food products with the use of the
scaling method [9].
2.5 Statistical Analysis
The differences between the mean values of
parameters examined (proximate composition, fatty
acid composition, sensory assessment) were calculated
using one-factor analysis of variance (repeated
measure) and SPSS 13.0.
3. Results and Discussion
3.1 Proximate Composition
The results of proximate analysis of kutum fish are
shown in Table 2.
In the present study, protein, lipid, ash and moisture
content of fresh kutum fish were found to be higher
than in frozen kutum fish for 12 months. While crude
protein, lipid, ash and moisture of fresh kutum fish are
21.8 0.02%, 3.21 0.01%, 1.29 0.01%, 75.9
0.04%, respectively, those values in frozen kutum
fish for 12 months were 19.9 0.01%, 1.91 0.04%,
1.83 0.05%, 72.3 0.04%, respectively.
Fish tissue possesses high nutritional value and is
therefore a particularly recommended directly
componential [10]. Lowering fish quality during frozen
storage has been attributed to undesirable changes
associated with lipid and proteins [11]. Changes in lipid
380 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets

Table 1 Sensory evaluation scores.

Score Color Odor Texture Flavor
Homogeneous, typical For cooked fish of Desirable, typical, no
5 typical, clearly Juiciness
certain perceptible foreign aftertastes
somewhat too soft or
4 Homogeneous Typical, poorly juicy
Desirable, typical, weak,
3 Somewhat Typical with slight foreign aftertaste poorly perceptible Semi-soft or hard Low juiciness
2 Un homogeneous Clearly perceptible Perceptible, foreign Soft or hard, fibrous Dry
1 Very un homogeneous Atypical, strong Strong foreign Very soft or very hard Very dry

Table 2 Proximate composition (dry weight basis) of kutum fish (Rutilus frisii kutum) subjected to different freezing periods
(MeanSE).
Freezing periods (months)
No. P-value
April June September December March
Ash 1.29 0.013 1.30 0.040 1.34 0.030 1.49 0.030 1.83 0.050 0.008**
Fat 3.21 0.01 2.43 0.01 2.52 0.01 2.30 0.02 1.91 0.04 0.03*
Protein 21.8 0.02 21.2 0.02 20.5 0.05 19.9 0.08 19.9 0.01 0.08
Moisture 75.9 0.04 75.2 0.01 74.5 0.01 73.6 0.04 72.3 0.04 0.03*
* The values are significantly different at P < 0.05. ** The values are significantly different at P < 0.01.

occur through hydrolysis and oxidation mechanism.
Fish muscle proteins can undergo denaturation during
frozen storage due to formation and accretion of ice
crystals resulting in dehydration [12]. Protein
aggregation in frozen fish depends on storage
temperature, temperature fluctuation, moisture changes,
storage time and enzymatic degradation [13]. The
effect of storage time on oil quality of horse mackerel,
shad, grad fish and golden mullet at-18 C in 150 days,
showed the highest oxidative deterioration that was
observed in shad oils [14]. A lipid analysis enabled the
classification and quantitative determination of 26 fatty
acids as well as the sum of SFA (Saturated Fatty Acids),
MUFA (Mono Unsaturated Fatty Acids ), PUFA (Poly
Unsaturated Fatty Acid), n-3 acids, n-6 acids and
n-3/n-6 ratio and changes in fatty acids (%total fatty
acids) of kutum fillet during frozen storage shown in
Table 3.
The fatty acid compositions ranged from 21.63% to
22.17% SFA (Saturated Fatty Acid), 37.83%-77.81%
MUFAs (Mono Unsaturated Fatty Acid),
36.11%-11.38% PUFAs (Poly Unsaturated fatty Acid)
and 74.09%-58.79% UFA (Unsaturated Fatty Acids).
Among them, those occurring in the highest
proportions in fresh samples were oleic acid (C18:1c,
20.57%), palmitic acid (C16:0, 13.91%) and DHA
(C22:6n3, 15.07%). These results are in agreement
with previous studies on FA of other species [5, 15-18].
The total saturated fatty acid content of lipids was
21.63% in fresh fillets and 22.17% in frozen fillets for
fish samples were stored for 12 months. Palmitic acid
(C16:0) was the primary saturated fatty acid,
contributing approximately 13.91% to the total
saturated fatty acid content of the lipids in fresh
samples. Similar results for Sander luciperca [5] and
Spratus aurita [15] (Boops boops); (Mugil cephalus);
(Sole solea) [3] and (Paralithodes camtschaticus,
Paralithodes platypus, Chionoecetes opilio,
Chionoecetes angulatus and Chionoecetes japonicus)
[19], (sparus sparus) [17] and green crab (Carcinus
mediterraneus) from the Tunisian mediterranean
coasts [20]. Palmitic acid was found to be the main
SFA, and was seasonal dependent with a mean value of
30.5% for E. aeneus, 27.9% for C. taeniops, and 20.9%
for S. scriba [18]. Palmitic acid was the most abundant
fatty acid in O. mossambica, C. carpio and L. rohita
ranging from 32% to 46% [21]. Palmitic acid was the
major SFA (14.6%-16.6%) and oleic acid was
identified as the major MUFA (15.1%-20.3%) for
Cyprinus carpio in Beysehir Lake all seasons [22].
 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate 381
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets

Table 3 Changes in fatty acids composition in kutum fish (Rutilus frisii kutum) during frozen storage at -18 C (%total fatty
acids).
Freezing periods (months)
Fatty acids Fresh fillet
April June September December March
C12:0 0.04 0.03 0.05 0.02 0.00 - - -
C13:0 0.01 0.01 0.00 0.01 0.00 - - -
C14:0 1.56 1.43 0.010 1.21 0.16 1.57 0.1 2.13 0.01 2.1 0.00
C15:0 1.02 0.66 0.01 0.72 0.03 0.78 0.1 0.83 0.13 0.80 0.00
C16:0 13.91 13.63 0.07 12.00 0.41 13.42 1.31 14.33 1.08 14.68 0.01
C17:0 0.72 0.65 0.03 0.76 0.03 0.82 0.02 0.80 0.08 0.74 0.02
C18:0 3.62 3.87 0.10 3.75 0.07 4.13 0.04 3.79 0.04 3.61 0.04
C20:0 0.07 0.11 0.02 0.10 0.05 0.12 0.03 0.17 0.08 0.12 0.00
C22:0 0.06 0.04 0.00 0.02 0.01 0.17 0.02 0.24 0.06 -
C24:0 0.65 0.38 0.01 0.39 0.13 0.37 0.19 0.11 0.01 -
C14:1 0.92 0.20 0.01 0.16 0.03 0.18 0.04 0.22 0.07 0.16 0.00
C16:1 11.51 10.85 0.62 9.29 0.73 10.83 1.21 11.68 1.21 11.69 0.02
C17:1 1.51 1.51 0.08 1.80 0.09 1.50 0.40 1.55 0.35 0.64 0.05
C18:1c 20.57 25.23 1.01 27.61 2.21 32.00 1.42 28.51 3.67 30.41 0.00
C20:1c 0.68 1.50 0.13 1.64 0.21 2.22 0.25 2.83 0.45 2.36 0.00
C22:1c 0.44 0.68 0.01 0.68 0.04 0.53 0.09 0.78 0.10 -
C24:1c 1.87 1.77 0.04 1.63 0.01 1.30 0.22 1.16 0.15 0.09 0.00
C18:2c 1.7 1.79 0.28 1.98 0.08 2.39 0.16 2.92 0.07 1.18 0.06
C18:3n-3 1.46 1.82 0.01 1.01 0.32 0.58 0.37 0.93 0.52 0.93 0.00
C18:4n-3 0.3 0.77 0.03 0.95 0.14 0.19 0.08 0.11 0.00 -
C20:2 0.15 0.11 0.05 0.23 0.11 0.34 0.04 0.33 0.10 0.44 0.00
C20:4n-6 6.98 3.83 0.18 3.61 7.99 2.45 1.02 1.58 0.19 0.17 0.00
C20:5n-3 7.69 6.27 0.13 4.92 0.66 2.81 0.41 2.46 0.51 2.73 0.07
C21:5n-3 0.75 0.85 0.03 0.61 0.12 0.68 0.20 0.47 0.15 -
C22:2 0.63 0.83 0.05 1.04 0.16 0.67 0.14 0.64 0.18 -
C22:5n-3 1.3 0.70 0.08 0.76 0.16 0.50 0.16 0.45 0.12 -
C22:6n-3 15.07 14.73 2.19 14.05 2.53 6.58 0.15 5.89 3.45 5.93 0.16
Values in the same line with different letters are significantly different at P < 0.01.

Oleic acid (C18:1c) was identified as the primary
monounsaturated fatty acid in fresh samples and was
significantly (P < 0.05) higher in fresh samples than in
fish samples were stored 12 months. Similar results for
E. aeneus, C. taeniops and S. scriba [18]. The major
fatty acids identified as polyunsaturated fatty acids
were EPA, (C20:5n3) and DHA (C22:6n3). EPA is the
most important essential fatty acid of the n-3 series in
the human diet because it is the precursor to the
3-series eicosanoids. It was confirmed that the fatty
acid profile of fish muscle reflects the content of the
dietary lipid sources. Compared with freshwater fish,
marine fish have higher levels of PUFAs, especially
DHA and EPA. It was reported that DHA decreases the
concentration of low density lipoprotein cholesterol in
plasma [15]. These results are in agreement with
previous studies on other species showed. The wild
European perched (Perca fluviatilis L.) were
characterized by lower concentrations of (MUFA)
Mono Unsaturated fatty acid and higher concentrations
of PUFA (Poly unsaturated Fatty Aacid) [23], and high
level of UFA (Unsaturated Fatty Acids) were in
Tunisian swordfish (Xiphias gladius). More than 40%
of the muscular tissues are made up of polyunsaturated
fatty acids (PUFA) of the n-3 series (EPA + DHA).
The products contain more than 30% of
monounsaturated acids characterized by the fatty acids
of the n-9 series and particularly by oleic acid [24]. The
382 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets

results of this study showed that total n-3 fatty acid was (18:3n-3), eicosatetraenoic (20:4n-3), and EPA
higher than n-6 PUFA. When fatty acids compared to (22:5n-3). Freshwater fish normally contain more n-6
other species, it was observed that the percent PUFAs, whereas marine fish are rich in n-3 fatty acids,
composition of PUFA in kutum fish fillets (36.11%) especially DHA and EPA [5]. Table 3 showed that
was higher than the reported values of Boops, Mugil, during the frozen storage, the difference in unsaturated
Sole, Sardin and Pandora (27.5%, 24.8%, 33.6%, fatty acids, expect for C16:0, C18:1c, C22:6n-3 were
31.02% and 32%), respectively. On the other hand, the significant (P < 0.05). On the other hand, initial PUFA
MUFA in the in kutum fish fillets (37.88%) was higher composition was 36.11% that decreased to 11.38% at
than the reported values of Sparus, Boops, Mugil, the end of storage time. Decrease in PUFA may be due
Trachurus, Sole and Scorpeana (29.87%, 27%, 25.8%, to breaking down of fatty acid chains, especially long
13.2%, 17.2%, 17.3%), respectively [3]. The fatty acid chain PUFA and the compositions with five or six
compositions of freshwater fish from Lake Seyhan double bonds (Table 4).
(Turkey) consisted of 28.0%-34.6% saturated (SFA), This was confirmed by data obtained by Zmijewski
10.7%-22.7% monounsaturated (MUFAs) and and Kujawa [26] who found that, freshwater
23.2%-43.7% polyunsaturated acids (PUFAs). The carnivorous fish can be characterized by greater
proportions of n-3 PUFAs of seawater fish (ranging n-3/n-6 fatty acid ratio than phytophagous and
from 22.6% for waker to 44.2% for blue fish) were benthophagous cyprinoid fish (tench, ide, crucian carp,
higher than those of n-3 PUFAs of freshwater fish grass carp). Differences in fatty acids of marine and
(ranging from 11.5% for North African catfish to freshwater fishes should not only be considered with
28.4% for zander). However, the levels of n-6 PUFAs respect to species habitat but also based on their natural
of sea water fish (ranging from 0.43% for blue fish to diet, especially whether a species is herbivorous,
14.4% for sea bass) were lower than those of n-6 omnivorous or carnivorous [27]. Apart from that, size,
PUFAs of fresh water fish (ranging from 5.27% for age, reproductive status of fish, environmental
kutum to 16.8% for Tench). The results showed that conditions, and especially water temperature, influence
fatty acid profiles of most fresh water fish are basically lipid content and fatty acid composition of fish muscle
comparable to those of seawater fish as sources of to a lipid content and fatty acid composition of fish
PUFAs [25]. It was also observed that the proportion of muscle to a certain extent [28]. Fatty acid composition
these fatty acids changed significantly in frozen storage of common carp (Cyprinus carpio), Nile tilapia
and oleic acid (C18:1c), palmitic acid (C16:0) and (Oreochromis niloticus) and tambacu, a hybrid of
DHA (C22:6 n-3) were predominant. The tambaqui (Colossoma macropomum) and pacu
concentrations of individual n-3 acids include linolenic (Piaractus mesopotamicus), was analyzed fresh and
Table4 Changes in Fatty acids series in kutum fish (Rutilus frisii kutum) during frozen storage at -18 C (%total fatty acids).
Freezing periods (months)
Fatty acids series Fresh fillets
April June September December March
SFA 21.11 19.23 21.38 22.40 22.17 21.63
UFA 73.51 72.66 66.40 62.29 58.79 74.09
MUFA 41.34 43.16 45.45 45.66 45.75 37.83
HUFA 26.38 23.95 13.02 10.38 10.49 31.79
PUFA 78.08 40.60 26.30 24.97 19.35 48.73
n-3 22.82 20.07 9.61 8.61 8.58 24.22
n-6 5.62 5.59 4.93 4.50 1.35 8.68
SFA: Saturated Fatty Acid; UFA: Unsaturated Fatty Acid; MUFA: Monounsaturated Fatty Acid; HUFA: High unsaturated Fatty Acid;
PUFA: Poly Unsaturated Fatty Acid.
 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets
The duration of freezing month
Fig. 1 Sensory evaluation of frozen kutum fish stored at -18 C for 12 months.
after 15, 30 and 45 storage days at -20 C. Total lipid
content was 9.3 g/100 g in carp, 0.79 g/100 g in tilapia
and 1.3 g/100 g in tambacu with skin, with reductions
of about 63%, 39% and 71% in the fishes without skin,
respectively. The carp showed a high content of
monounsaturated fatty acids (about 50%). In tilapia,
palmitic and oleic acids were present in larger
proportion, 26.55% and 23.86%, respectively. In
tambacu, the fatty acid profile was 37% saturated 34%
MUFA and 21% PUFA. Fatty acid composition did not
present wide variations due to storage time and
preparation, indicating that the storage and cooking
methods used did not interfere in fatty acid
composition [29].
Pomfret (Pampus punctatissimus) oil saturated fatty
acids were the most abundant (48.3%). Palmitic acid
(16:0) was the dominant fatty acid, followed by oleic
acid (18:1), DHA (22:6n-3), myristic acid (14:0) and
stearic acid (18:0), with percentages of 30.5, 26.3, 12.2,
7.37 and 6.86, respectively. The ratio of n-3/n-6
polyunsaturated fatty acids (PUFAs) was 8.04; thus,
pomfret muscle is rich in n-3 PUFA [29].
A comparison of sensory properties in the
above-described study confirmed that, there were
changes in all the sensory storage. Frozen fish is better
achieved in the 180 days of storage. Deterioration
increases as the duration of storage increases (Fig. 1).
4. Conclusions
Seafood origin proteins and fatty acids play an
383
important part in the human diet. The fatty acids
PUFAs and HUFAs are crucial in terms of human
feeding physiology. These fatty acids contain five or
six double bonds in their body. The most important of
HUFAs and PUFAs are n-3 fatty acids, and are
particularly found in freshwater fish. This condition
can be regarded as an explanation for the fact that the
kutum fish (Rutilus frisii kutum) is richer in n-3 fatty
acids taking into consideration the fatty acid profile of
the fish. On the other hand, unsaturated fatty acids
contents of frozen kutum fish maintained their quality
during 12 months of storage.
References
[1] M.I. Gladyshev, N.N. Sushchik, G.A. Gubanenko, S.M.
Demirchieva, Effect of way cooking on content of
essential polyunsaturated fatty acids in muscle tissue of
humpback salmon (Oncorhynchus gorbuscha), Journal
Food Chemistry 96 (2006) 446-451.
[2] J. Mayneris-Perxachs, I. Bondia-Pons, L. Serra-Majem, A.
Castellote, M. Lpez-Sabater, Long-chain n-3 fatty acids
and classical cardiovascular disease risk factors among the
Catalan population, Journal Food Chemistry 119 (1)
(2010) 54-61.
[3] Y. Ozogul, F. Ozgoul, Fatty acid profiles of commercially
important fish species from the Mediterranean, Aegean
and Black seas, Journal Food Chemistry 100 (2007)
1636-1638.
[4] S.T. Arannilewa, S.O. Salawu, A. Sorungbe, Effect of
frozen period on the chemical, microbiological and
sensory quality of frozen tilapia fish (Sarotherodun
galiaenus), Journal Nutr Health 18 (2) (2006) 185-192.
[5] M. Celic, A. Diler, A comparison of the proximate
composition and fatty acid profiles of Zander (Sander
384 Influence of Frozen Storage Time on Unsaturated Fatty Acids and Proximate
Composition of Caspian Sea White Fish (Rutilus frisii kutum) Fillets
lucioperca) from two different region and climatic
condition, Journal Food Chemistry 92 (2005) 637-641.
[6] B.G. Begona, J. De Souse, Chemical changes and visual
appearance of albacore tuna as related to frozen storage,
Journal of Food Science 64 (1) (1999) 20-24.
[7] .G. Blight, W.J. Dyer, A rapid method of total lipid
extraction, Journal Biochemistry 37 (1959) 911-917.
[8] A.O.A.C., Official Methods of Analysis, 15th ed.,
Association of Official Analytical chemists, Arlington,
V.A., The Association, USA, 1990.
[9] ISO 8586-1, Sensory analysis-general guidance for the
selection, training and monitoring of assessors-part 1:
Selected assessors, 1993.
[10] S. Abourg, C. Pineiro, Biochemical changes and quality
chilled storage of farmed turbot (Psetta maxima), Journal
Food Chemistry 90 (3) (2004) 445-452.
[11] A. Huidobro, M. Tejado, Gilthhead sea bream (Sparus
sparus): Suitability for freezing and commercial
alternatives, Journal of the Science of Food Agriculture 84
(11) (2004) 1405-1410.
[12] S. Simoenidou, A. Govaris, Qualtiy assessment of seven
Mediterranean fish species during storage on ice, Journal
Food Research International 30 (7) (1998) 479-484.
[13] C. Bauchart, C. Chambon, Peptides in rainbow trout
(Onchorhynchus mykiss) muscle subjected to ice storage
and cooking, Journal Food Chemistry 100 (4) (2007)
1566-1572.
[14] G. Boran, H. Karaam, M. Boran, Changes in the quality
of fish oils due to storage temperature and time, Journal
Food Chemistry 98 (4) (2006) 693-698.
[15] A. Zuraini, M.N. Somchit, Fatty acid and amino acid
composition of three local Malaysian Channa spp. fish,
Journal Food Chemistry 97 (2006) 674-678.
[16] A. Mnari, I. Bouhle, Fatty acid in muscle and liver of
Tunisian Wild and farmed gilthead Sea bream (Sparus
aurata), Journal Food Chemistry 100 (2007) 1393-1397.
[17] S. Abourg, C. Pineiro, Biochemical changes and quality
chilled storage of farmed turbot (Psetta maxima), Journal
Food Chemistry 90 (3) (2004) 445-452.
[18] A. Ahmed Louly, El Kebir E. Gaydou, Muscle lipids and
fatty acid profiles of three edible fish from the Mauritanian
coast: Epinephelus aeneus, Cephalopholis taeniops and
Serranus scriba, Journal Food Chemistry 124 (1) (2011)
24-29.
[19] N.A. Latyshev, S.P. Kasyanov, V.I. Kharlamenko, V.I.
Svetashev, Lipids and of fatty acids of edible crabs of the
north-western Pacific, Journal Food Chemistry 116 (3)
(2009) 657-661.
[20] S. Cherif, F. Frikha, Y. Gargouri, N. Miled, Fatty acid
composition of green crab (Carcinus mediterraneus) from
the Tunisian Mediterranean coasts, Journal Food
Chemistry 111 (4) (2008) 930-933.
[21] F. Jabeen, A. Chaudhry, Chemical compositions and fatty
acid profiles of three freshwater fish species, Journal Food
Chemistry 125 (37) ( 2010) 991-996.
[22] G.O. Guler, B. Kiztanir, A. Aktumsek, O.B. Citil, H.
Ozparlak, Determination of the seasonal changes on total
fatty acid composition and 3/6 ratios of carp (Cyprinus
carpio L.) muscle lipids in Beysehir Lake (Turkey),
Journal Food Chemistry 108 (2) (2008) 991-996.
[23] B. Jankowska, Z. Zak, T. mijewski, M. Szczepkowski,
Fatty acid profile of muscles, liver and mesenteric fat in
wild and reared perch (Perca fluviatilis L.), Journal Food
Chemistry 118 (3) (2010) 764-768.
[24] M.A. Ben Smida, B. Marzouk, M. El Cafsi, The
composition of fatty acids in the tissues of Tunisian
swordfish (Xiphias gladius), Journal Food Chemistry 115
(2) (2009) 522-528.
[25] Y. Ozogul, F. Ozgoul, S. Alagoz, Fatty acid profiles and
fat contents of commercially important seawater and
freshwater fish species of Turkey: A comparative study,
Journal Food Chemistry 103 (1) (2006) 217-223.
[26] . Zmijewski, R. Kujawa, Slaughter yield, proximate and
fatty acid composition and sensory properties of rapfen
(Aspius aspius), with tissue of bream (Abramis brama L.),
and pike (Esox lucius L.), Journal of Food Composition
and Analysis 19 (2006) 176-181.
[27] F. Badii, N. Howell, A comparison of biochemical changes
in cod (Gadus morhua) and haddock (Melanogrammus
aeglefinus) fillets during frozen storage, Journal of the
Science of Food and Agriculture 82 (2001) 87-97.
[28] E.A. Bryhni, Poly unsaturated fat and fish oil in diets for
growing-fishing pigs: Effects on fatty acid composition
and meat, fat and sausae quality, Journal Meat Science 62
(1) (2002) 1-8.
[29] F. Ferreira de Castro, H. Pinheiro SantAna, F. Campos,
M. Brunoro Costa, M. Coelho Silva, A. Salaro, S. Castro
Franceschini, Fatty acid composition of three freshwater
fishes under different storage and cooking processes,
Journal Food.
Journal of Agricultural Science and Technology B 2 (2012) 385-390
Earlier title: Journal of Agricultural Science and Technology, ISSN 1939-1250
D DAVID PUBLISHING

The Influence of Low Frequency Electromagnetic Field

(LF EMF) on the Agricultural Crops Seeds Germination

Mihail Gennadevich Barishev1,2, Stepan Sergeevich Dzhimak1,2, Gennadij Ivanovich Kasjanov3 and Denis
Igorevich Sashhkov1
1. University of Kuban State, Krasnodar Region 350040, Russian Federation
2. Southern Scientific Center of Russian Academy of Sciences, Rostov Region 344006, Russian Federation
3. Technical University of Kuban State, Krasnodar Region 350072, Russian Federation

Received: June 7, 2011 / Published: March 20, 2012.

Abstract: This research is about changing rate of seeds germination by low frequencies electromagnetic field. The method of
resonant frequencies to increase seeds germination were developed and patented by authors of this research. The idea of the means is
to research a technique to changing a capacitor component of full resistance for extracts of wheat and sunflower seeds under the
influence of low frequency electromagnetic field (LF EMF). After the green sprouting experimentin cups of Petri, it was found that
most frequencies found by this method are good, and the best results increasing germination of wheat of a grade Krasnodar 99 were
after treatment with 31 Hz, for sunflower of a grade Buzuluk 16 Hz.

Key words: Augmentation germination seeds, low frequencies, electromagnetic processing, extracts from seeds.

1. Introduction mechanisms of influence EMF on biological systems

recognize that one of the processes most sensitive to
One of the major anthropogenous factors is an
external influences transitions of various fibers, in
electromagnetic field of technogenic sources. Most
particular peripheric, from the condition bound on
acutely this problem has risen in the middle of 20
membranes on water medium are. Similar
centuries. Now the quantity of the technogenic sources
unidirectional processes occur at certain stages of an
generating electromagnetic field (EMF) and
exit of seeds from a rest condition. Such transitions of
negatively influencing doesnt decrease for
fibers, because of growth of number of degrees of
environment, but only increases. It is bound to
freedom for albuminous groups in water medium and
constant development of a radio communication, a
according to entropy of system, should be bound to
radar-location, radio navigation, television and other
small change of free energy. They can be caused local
communication media. Within last 30 years as a result
changes sensitive to influence EMF either ionic
of works of some researchers high sensitivity of
force, or concentration of ions of Ca2+, and the
biological systems to action on the electromagnetic
modeling data show that effects of EMF in the field of
field of low frequencies (EMF LF) weak intensitya
low frequencies (from 0.1 to 102 Hz) can be
low-frequency range is established. Using EMF LF
essentially strengthened at the expense of nonlinear
for regulation of activity of biological objects is a
processes in under membrane a layer.
perspective direction. By treating of probable
Presence of own fluctuations of conductivity in the
Corresponding author: Stepan Sergeevich Dzhimak, Ph.D., water, lying in this range can be one of the possible
research fields: low frequencies electromagnetic field, NMR, reasons of the inadequate answer of biological objects
water structure, deuterium depleted water. E-mail:
jimack@mail.ru. on action weak EMF low frequencies. The majority of
386 The Influence of Low Frequency Electromagnetic Field (LF EMF)
on the Agricultural Crops Seeds Germination
biopolymers function in water medium, interaction of
monomers making them define a spatial configuration
of macromolecules. In this connection effects of
change of structure of water near to molecules of the
dissolved substances essentially should be reflected in
a configuration of macromolecules, in processes of
active transport of large molecules through a
membrane by means of permiaz, and also passive
transport of ions through a membrane [1].
Rising of productivity of agricultural crops and
improvement of quality of production, is one of actual
problems of agriculture. Now researches of influence
EMF LF on agricultural crops for the purpose of rising
of productivity and improvement of quality of
production are perspective. In Krasnodar region one of
the most widely grown up cultures are wheat and
sunflower. Therefore we had been conducted
researches of influence EMF LF on these cultures.
2. Experiment
LF made processing of seeds EMF LF by means of
the equipment represented in Fig. 1. For processing a
frequency range from 12 Hz to 40 Hz, magnetic
induction EMF LF made 0.2 m. Time of processing
of seeds EMF made LF of 20 minutes. Definition of
resonant frequencies for processing of seeds carried
out by a technique of measurement of a capacitor
component of full resistance of an extracting solution
under the influence of EMF LF [2].
For researches an extract from seeds placed in an
electric cell with the electrodes covered with platinum
common people. A cell connected to impedancemeter.
Simultaneously made constant control of temperature
of investigated extracts by means of the electron
thermometer to within 0.1 C. Definitions of resonant
frequencies of investigated systems spent with the
help impedancemeter M 507. Peaks of maximums
correspond to stimulating frequency.
In researches sunflower seeds of a grade Buzuluk,
wheat seeds of a grade Krasnodar 99 were used.
Measured energy of germination seeds, for sunflower
1: oscillator G3 118, 2: a frequency meter, 3: an ultrasonic head,
4: capacity for loading of investigated biosystems, 5: the
shielding chamber.
Fig. 1 The block diagram of the device for processing of
investigated biosystems and water of EMF KNC.
also measured lengths of sprouts and hipokotiley
Sampling carried out according to GOST 12036-85
[1], chose seeds which didnt have mechanical
damages, and the sizes of seeds lay within 1.4-1.6 sm;
Rejected the seeds possessing abnormal coloring,
mouldy, with the damaged cotyledons, seeds without a
germ, the sprouted seeds. Seeds landed in cups of
Petri, for germination used the distilled water specific
resistance of 300 k Omega/sm.
Definition of energy of germination seeds
sunflower and wheat seeds made in accordance with
GOST 12038-84 [1]. For research used seeds of
sunflower of a grade Buzuluk, seeds of wheat of a
grade Krasnodar 99.
3. Results and Discussion
From spent before researches on influence EMF LF
on biological objects, it has been established that for
wheat and sunflower seeds exist both optimum zones,
and a zone pessimum on intensity of a field [1].
Researches spent in an optimum zone, intensity EMF
LF 400 /m.
In Fig. 2, dependence of a capacitor component of
full resistance of an extract from seeds from frequency
is presented. Apparently from Fig. 2, at preseeding
processing of seeds of sunflower of a grade Buzuluk
in a stimulating regimen, frequencies 16, 17, 18, 19,
22 and 32 Hz can be used.
 The Influence of Low Frequency Electromagnetic Field (LF EMF)
on the Agricultural Crops Seeds Germination
C,
350
16
22
300
19
18
250 12
14
17
200 21
13
150
23
100 15
20
50
10 15 20
Fig. 2 Dependence a capacitor component of full resistance for extracts of wheat and sunflower seeds under the influence of
low frequency electromagnetic field (LF EMF). = 400 /.
Results of research of energy of germination of seeds
of sunflower of a grade Buzuluk, processed EMF LF on
certain resonant frequencies are presented in Fig. 3.
The analysis of the presented data shows that the
greatest effect on energy of germination of sunflower
seeds is observed at frequency f = 16 Hz. Thus skilled
value surpassed control on 20%, the maximum error
didnt exceed 6% (at reliability = 0.95). Along with
research on energy of germination by us estimated
germination seeds (Fig. 4).
Apparently from the data presented on Fig. 4,
maximum germination of sunflower seeds are
observed at their processing EMF of the same
frequency, as at definition of a maximum of energy of
germination, 16 Hz. At influence EMF with this
frequency germination pre-production model has
exceeded control also on 20%. Thus, it is
experimentally defined that the most suitable mode of
exposure of EMF KNC on sunflower seeds of a grade
Buzuluk is f = 16 Hz, at intensity of magnetic field
= 400 /m, and influence time t = 20 minutes. It is
under existing conditions observed maximum in
comparison with control germination and energy of
germination of seeds, and also absence on seeds
moldy microfloras.
However as have shown results of laboratory
387
32
33
37
40
35
38
27 30 36
39
25
24 2829
31
26 34
25 30 35 40 45
f,
trialsat research of influence of efficiency of
influence of EMF KNC on initial stages of
germination of seeds not always enough measurement
of such parameters, as energy of germination and
germination seeds. Often enough at processing of
seeds of EMF KNC identical value of the given
parameters are observed at processing of seeds by
different frequencies whereas the length of sprouts
and roots (hipokotiley) seeds considerably vary
depending on frequency EMF at which there is a
processing. Therefore we had been spent
measurements of length of sprouts and roots of seeds
and statistical processing of the received data by
means of an estimation of a difference of selective
averages by t-criterion for reliability of 95%.
Apparently from the Table 1, processing of seeds of
sunflower with frequency of 16 Hz EMF leads to the
greatest authentic change of length of sprouts and
roots, i.e. 16 Hz-are optimum frequency, for
stimulation of seeds of sunflower of a grade Buzuluk.
At processing of seeds with frequency of 22 Hz
difference of length of roots from control has made
32%, however difference of average length of sprouts
from control has made only 10% and value
germination seeds at processing of EMF KNC with
the given frequency has made only 83%.
388 The Influence of Low Frequency Electromagnetic Field (LF EMF)
on the Agricultural Crops Seeds Germination

E,%
100

95

90

85

80

75

70
0 16 17 18 19 22 32
f,

Fig. 3 Dependence a capacitor component of full resistance for extracts sunflower seeds, a grade Buzuluk, under the
influence of low frequency electromagnetic field (LF EMF), t = 20 minutes, = 400 /m.

V, %
100

95

90

85

80

75
0 16 17 18 19 22 32
f,

Fig. 4 Dependence germination seeds of sunflower of a grade Buzuluk from frequency EMF LF (t = 20 minutes, = 400
/m).

Table 1 Difference a capacitor component of full resistance for extracts sunflower seeds a grade Buzuluk from control at a
stage germination and value of t-parameter.
Frequency of Theoretical
16 17 18 19 22 32
processing EMF, Hz value of
Rostock and root Rost Root Rost Root Rost Root Rost Root Rost Root Rost Root t-criterion,
Percent from control 21 32 15 28 14 30 18 29 10 32 19 25 for 0.95
Value of t-criterion 2.58 4.98 1.82 4.06 1.63 4.36 1.69 4.17 1.2 5.23 2.04 3.64 1.98

Results of definition of resonant frequenciesby resistance component of an extracting solution under

means of a measurement technique of a capacitor the influence of EMF LF, are presented in Fig. 5
processing of wheat seeds, a grade Krasnodar 99, full Apparently from Fig. 5, for preseeding processing of
 The Influence of Low Frequency Electromagnetic Field (LF EMF) 389
on the Agricultural Crops Seeds Germination

,
250

200 31
16 26
19

150 21 39
32 36
29
37
17 23
100
12 24 27
30
18
15 25 35 38
11 13 14 33 34
50
10 22
20 28

0
10 15 20 25 30 35 40

f,

Fig. 5 Definition of resonant frequencies for processing EMF LF of wheat of a grade Krasnodar 99 at = 400 /m.

seeds of wheat of a grade Krasnodar 99 in a stimulating
regimen, frequencies 16, 19, 26 and 31 Hz can be used.
Results of research of energy of germination grade
wheat Krasnodar 99 are presented in Figs. 6 and 7.
The analysis of the received dependences shows
that the greatest effect of energy dependence of
germination wheat seeds from frequency under the
influence an electromagnetic field is observed on
frequency f = 31 Hz, thus skilled value surpassed
control on 15%, the maximum error didnt exceed
6% (at reliability = 0.95).
Here maximum was observed on the same
frequency for maximum of energy of germination (31
Hz). The influence of EMF with this frequency
germination preproduction model has exceeded control
also on 19%. Thus, it is experimentally established that
the most suitable mode of exposure of EMF KNC on
seeds of wheat of a grade Krasnodar 99 is f = 31 Hz, at
intensity of magnetic field = 400 /m, influence
time t = 20 minutes. At the given parameters of
processing it is observed maximum in comparison with
control germination both energy of germination of
seeds and absence of a mushroom microflora.
Germination seeds of wheat and sunflower under
the influence of a magnetic field, it is possible to
explain the change that water in seeds changes the
structural organization that causes changes osmotic
pressure in cells. In case of stimulating influence
(leading to augmentation germination), osmotic
pressure in cells increases that leads to augmentation
of degree of a stretching of membranes and reduction
of communication of peripheric fibers with a
membrane. There is liberation of larger quantity of
fibers, i.e., process of transition of fibers of the bound
condition in dissolved, and start with their help of the
further processes necessary for germination of seeds is
accelerated. In a case of influence EMF leading to
reduction germination of seeds, there is a return
process [3, 4]. The density and viscosity of water is
enlarged, osmotic pressuredecreases. As a result,
the probability of liberation of the peripheric fibers
bound to a membrane, decreases. It is possible that
electromagnetic field of low frequencies reorientation
molecules of water in a blanket of membranes [5, 6].
4. Conclusions
As a result of the spent researches communication
of change of resonant frequency EMF and energies of
germination and germination seeds is visible.
Resonant frequencies to increase germination seeds
were found by means of a technique of research of
change of a capacitor component of full resistance of
an extract from seeds of wheat and sunflower under
the influence of an electromagnetic field of low
390 The Influence of Low Frequency Electromagnetic Field (LF EMF)
on the Agricultural Crops Seeds Germination

, %
90

85

80

75

70

65

60

55

50
0 16 19 26 31
f,

Fig. 6 Dependence in energy germination of wheat seeds of a grade Krasnodar 99 from frequency EMF LF, t = 20 minutes,
= 400 /m.

V, %
100

95
90

85
80

75
70

65
60

55
50
0 16 19 26 31
f,

Fig. 7 Dependence germination seeds of wheat of a grade Krasnodar 99 from frequency EMF LF, t = 20 minutes, = 400 /m.

frequencies. For wheat there are 16, 19, 26, 31 Hz,
and for sunflower 16, 17, 18, 19, 22, 32 Hz. After
experimental greens prouting in cups of Petri, it was
found that the best frequency to increase germination
of wheat of a grade Krasnodar 99 is 31 Hz, for
sunflower of a grade Buzuluk 16 Hz.
References
[1] M.G. Barishev, N.N. Kulikova, S.S. Dzhimak, et al.,
Influence of a Low-Frequency Electromagnetic Field on
Biological Systems, SSC the Russian Academy of
Sciences Press, Rostov-on-Don, 2008, p. 288.
[2] M.G. Baryshev, O.V. Evdokimova, L.V. Lomakina, S.S.
Dzhimak, Equipment for Adjustment of Activity of
Biological Objects, Patent of the Russian Federation No.
67381, 2007.
[3] S.I. ksenov, Water and Its Role in Regulation of
Biological Processes, Institute of Computer Researches
Press, Izhevsk, 2004, p. 212.
[4] S.I. Aksenov, T.J. Grunina, .. Gorjachev, About
stimulation and inhibition mechanisms at germination of
seeds of wheat in an electromagnetic field of ultralow
frequency, Biophysics 52 (2007) 332-338.
[5] A. Goldsworthy, H. Whitney, E. Morris, Biological
effects of physically conditioned water, Wat. Res. 33
(1999) 1618-1626.
[6] J. Teissie, Biophysical effects of electric fields on
membrane water interfaces: A mini review, Eur. Biophys.
J. Biophys. Lett. 36 (2007) 967-972.