Moringa oleifera as an Alternative

Fodder for Dairy Cows in Nicaragua

Bryan Mendieta-Araica
Faculty of Veterinary Medicine and Animal Science
Department of Animal Nutrition and Management
Uppsala
Doctoral Thesis
Swedish University of Agricultural Sciences
Uppsala 2011

Acta Universitatis agriculturae Sueciae
2011:34

ISBN 978-91-576-7569-9
ISSN 1652-6880
2011 Bryan Mendieta-Araica, Uppsala
Print: SLU Service/Repro, Uppsala 2011
Cover: Blooming Moringa oleifera tree
(Photo: B. Mendieta-Araica)

Moringa oleifera as an Alternative Fodder for Dairy Cows in
Nicaragua
Abstract
The four studies comprising this thesis characterised Moringa oleifera as a fodder for
dairy cows under dry tropical conditions in Nicaragua. An agronomy study
examined, two planting densities (D
1
=100,000 and D
2
=167,000 plants ha
-1
) and four
fertilisation levels (N
1
=0, N
2
=261, N
3
=521 and N
4
=782 kg N ha
-1
). The D
2
density
gave significantly higher yields of total dry matter ha
-1
(TDMY) and fine fraction dry
matter ha
-1
(FFDM) compared with D
1
. There were significant interactions between
fertilisation level and the variables year and cut with regard to TDMY and FFDM.
However, fertilisation levels N
3
and N
4
gave the highest yield in both years and
among all cuts.
A study on Moringa leaf meal (MLM), as a protein source in concentrates to dairy
cows found no significant difference in milk production when comparing isocaloric
and isoproteinic concentrates with or without MLM. In an ensiling experiment,
Moringa was ensiled alone with 10 g kg
-1
fresh matter (FM) molasses and compared
with several mixtures with Elephant grass and sugar cane. Pure Moringa biomass
produced silage with a higher crude protein (CP) content and had a favourable
effect on silage pH, with higher lactic acid concentrations, but the presence of
Moringa decreased time to spoilage by 67 h (22%) compared with the Elephant
grass silages.
Feeding Moringa as the sole roughage, either fresh or ensiled, compared with
feeding Elephant grass resulted in higher digestibility of both CP and fibre but milk
yield did not differ (13.7 kg cow day
-1
). No differences in milk composition were
found between treatments but when fresh Moringa was fed a grassy flavour and
aroma was detected in the milk.
In conclusion, to maintain high biomass yield of Moringa over time, the best
planting density-fertiliser combination was D
2
and N
3
. MLM can successfully replace
commercial concentrate ingredients for dairy cows. Furthermore, Moringa ensiled
alone, with only 10 or 50 g kg
-1
FM molasses added, produces good quality silage
that can be fed to dairy cows in large quantities while maintaining the same milk
production level and milk quality as for cows fed conventional roughages.

Keywords: Moringa oleifera, dairy cows, milk yield, milk composition, organoleptic
characteristics, silage, leaf meal, biomass yield, planting density, fertilisation levels.

Authors address: Bryan Mendieta-Araica, Departamento de Sistemas Integrales de
Produccin Animal, Facultad de Ciencia Animal, Universidad Nacional Agraria,
UNA, Km 12 carretera Norte, Managua, Nicaragua, Apdo 453. E-mail:
bryan.mendieta@una.edu.ni


Dedication
To the memory of my
Mother Isabel Araica, the most wonderful woman ever;
And to my beloved family,
My kids: Brianna Isabella
Bryan Giancarlo
Bryan Jeanfranco
And my wife Ivania.


If I have seen further it is only by standing on the shoulders of giants.
Sir Isaac Newton


Contents
List of Publications 7
Abbreviations 9
1 Introduction 11
2 Aims 15
3 Summary of Materials and Methods 17
3.1 Treatments 17
3.2 Locations 18
3.3 Planting Moringa 18
3.3.1 Soil preparation and sowing 18
3.3.2 Fertilisation 18
3.4 Feeds 19
3.4.1 Moringa leaf meal (MLM) 19
3.4.2 Concentrates 19
3.4.3 Ensiling 19
3.4.4 Fresh forages 20
3.5 Animal Management 20
3.6 Sampling 21
3.6.1 Biomass 21
3.6.2 Feed and faeces 21
3.6.3 Milk 22
3.7 Chemical Analysis 22
3.8 Organoleptic Characteristics 23
3.9 Experimental Design and Statistical Analysis 24
4 Summary of the Results 27
4.1 Moringa as a crop (Paper I) 27
4.2 Moringa leaf meal in concentrates (Paper II) 28
4.3 Moringa as silage (Paper III) 29
4.4 Moringa as roughage (Paper IV). 29
5 General Discussion 31
5.1 Biomass production 31
5.2 Characterisation of Moringa as a feedstuff 33
5.2.1 Chemical composition of Moringa foliage 33
5.2.2 Chemical composition of Moringa silage 37

5.2.3 Labour requirement for silage making and leaf meal production 38
5.2.4 Practical implications of using different feed products from
Moringa 39
5.3 Effect of Moringa on milk yield and milk composition 40
5.4 Effect of Moringa on organoleptic characteristics of milk 41
5.4.1 Flavour 41
5.4.2 Aroma 42
5.4.3 Colour and appearance 43
6 Main Findings and Conclusions 45
7 Future research 47
References 49
Acknowledgements 57

7
List of Publications
This thesis is based on the work contained in the following papers, referred
to by Roman numerals in the text:
I Mendieta-Araica, B., Sprndly, E., Reyes-Snchez, N., Salmern-
Miranda, F., Halling, M., (2011). Biomass production and chemical
composition of Moringa oleifera under different planting densities and
levels of nitrogen fertilization (Submitted to Agroforestry Systems).
II Mendieta-Araica, B., Sprndly, R., Reyes-Snchez, N., Sprndly, E.,
(2011). Moringa (Moringa oleifera) leaf meal as a source of protein in
locally produced concentrates for dairy cows fed low protein diets in
tropical areas. Livestock Science 137, 10-17.
III Mendieta-Araica, B., Sprndly, E., Reyes-Snchez, N., Norell, L.,
Sprndly, R., (2009). Silage quality when Moringa oleifera is ensiled in
mixtures with Elephant grass, sugar cane and molasses. Grass and Forage
Science 64, 364-373.
IV Mendieta-Araica, B., Sprndly, E., Reyes-Snchez, N., Sprndly, R.,
(2011). Feeding Moringa oleifera fresh or ensiled to dairy cows- Effects on
milk yield and milk flavor. Tropical Animal Health and Production 43,
1039-1047.

Papers II-IV are reproduced with the permission of the publishers.
8


9
Abbreviations
ADF Acid detergent fibre
AOAC
APHA
BCN
C
CFU
CP
CT
CV
DM
ECM
EG+C
FAO
FFDM
FM
GLM
GR
H
IU
LAB
MAGFR
ME
MF
MLM
MPN
MS
NDF
NRC
Association of official analytical chemists
American public health association
Central Bank of Nicaragua
Degrees Celsius
Colony-forming units
Crude protein
Tissue culture
Cultivar
Dry matter
Energy corrected milk
Elephant grass + concentrate
Food and agriculture organization
Fine fraction dry matter
Fresh matter
General linear model
Growth rate
Plant height
International units
Lactic acid bacteria
Ministry of Agriculture and Forestry
Metabolisable energy
Moringa fresh
Moringa leaf meal
Most probable number
Moringa silage
Neutral detergent fibre
National research council
10
OM
SAS
SBM
SIDA
TDN
TDMY
UNA
USD
USDA
WSC

Organic matter
Statistical analysis system
Soybean meal
Swedish international development agency
Total digestible nutrients
Total yield of dry matter
National University of Agriculture
United States dollars
United States Department of Agriculture
Water soluble carbohydrates









11
1 Introduction
Historically, the livestock sector has played an important role in Nicaragua,
and the significant social and economic importance of the sector remains.
The livestock sector accounts for 35% of agricultural gross domestic product
and it employs about 200,000 people of an economically active population
of 641,000 persons (BCN, 2009). At the farm level, livestock is seen as a
source of wealth and insurance and has a high cultural significance
establishing the status of the farmer.
Early in the 1990s, the dairy sector in Nicaragua started an extensive and
rapid growth process. Dairy production grew from about 30 million litres of
milk in 1990 to 730 million litres in 2009 (BCN, 2009; MAGFOR, 2009).
Exports of dairy products, which generated revenue of USD 18.4 million in
2001, had grown to USD 116.1 million by 2009 (BCN, 2009). In less than
two decades, Nicaragua became a country with net dairy exports.
Nicaraguan farmers currently face opportunities and challenges. Nicaragua is
well positioned to produce and expand dairy production for local and export
markets because it has over three million head of cattle, the biggest herd in
Central America (FAO, 2006) and good processing capacity. However, the
challenge lies in many issues such as low milk yield, long calving intervals,
low calving rate, underfeeding due to limitations in the quality and quantity
of feed, mainly during the dry season, and high vulnerability to climate
change. Furthermore, the livestock systems are mainly pasture-based and
have been progressively moving towards marginal areas with less productive
capacity (Holmann et al., 2004; Mendieta-Araica et al., 2000; Kaimowitz,
1995).
Livestock systems have been cited as the major cause of environmental
degradation such as deforestation, soil degradation, production of
greenhouse gas emissions and of rural poverty (Belli et al., 2009; Mauricio et
al., 2008; Steinfeld et al., 2006). Therefore sustainable alternatives in animal
production need to be developed and adopted to reduce the negative impact

12
of livestock on the environment. Furthermore, there is an urgent need to
enhance the efficiency of natural resource use in livestock production in
order to overcome the strong seasonal effect on production and
consequently improve the standard of living for farmers. There are many
options that can be used to deal with the problems described above.
However, the strategy that is finally adopted should take into account not
only the technical aspect of the problem, but also the economic and social
aspects. Interesting options can be found of silvopastoral systems that
combine pasture with the production of shrubs and trees. According to
some authors (Carvalho et al., 2001; Snchez and Speed, 1999; Castro et al.,
1998) these agro forestry systems offer sustainable management techniques
for animal production in the tropics and provide a number of supplementary
economic, social and environmental benefits.
In Nicaragua, forage trees such as Leucaena leucocephala, Gliricidia sepium,
Erytrhina spp. and Guazuma ulmifolia have been studied for a long time and
their role in silvopastoral systems and effects on animal production are well
documented (Mosquera-Losada et al., 2005; Durr, 1992; Beer, 1989).
However, many of these forage trees have not been widely used because
they often contain anti-nutritional compounds that have deleterious effects
on animal yield (Ghosh et al., 2007; Hammond, 1995). Therefore,
researchers have increasingly been paying attention to Moringa (Moringa
oleifera Lam synonym: M. pterygosperma Gaertn., M. moringa Mills), which is a
widespread, drought-tolerant tree with negligible amounts of tannins,
trypsin and amylase inhibitors (Gidamis et al., 2003; Makkar and Becker,
1997; Becker, 1995). It can have a total dry matter (DM) yield up to 24 ton
ha
-1
year
-1
(Reyes-Snchez et al., 2006a) and has a crude protein (CP) content
in fresh leaves varying from 193 to 264 g kg
-1
DM.
Moringa fresh foliage has been included into the diet of different animals.
Positive effects on feeding behaviour in goats (Manh et al., 2005), growth
rate in sheep (Ben Salem and Makkar, 2009) and milk yield in dual purpose
cows (Reyes-Snchez et al., 2006b) have been reported. Moringa can be also
dried and used in the form of Moringa leaf meal (MLM). Promising results
have been obtained on inclusion of MLM into the diet of fish (Richter et al.,
2003), sheep (Murro et al., 2003), laying hens (Kakengi et al., 2007) and
cross-bred dairy cows (Sarwatt et al., 2004). However, reports on feeding
MLM to dairy cows are still few. No reports have been found about Moringa
silage and its use, or the use of fresh Moringa foliage as the entire roughage
component in the dairy cow ration. Interestingly, Moringa has been
reported to be a valuable component in human food due to its adequate
amino acid profile and CP content, its high level of vitamin A and its low
13
level of anti nutritional compounds (Snchez-Machado et al., 2010;
Anhwange et al., 2004).
In order to be a suitable source of fodder, Moringa should be able to
produce and maintain high biomass yields over the years. However, under
dry tropical forest conditions and without fertilisation, an interannual
reduction in yield was reported by Reyes-Snchez et al. (2006a). Only a few
studies on the mineral nutrition of this plant have been performed, all of
them under laboratory conditions (Dash and Gupta, 2009; Oliveira et al.,
2009; Pamo et al., 2005). Therefore studies of fertilisation strategies in
Moringa are required.
14





15
2 Aims
The overall aim of this thesis was to gain further knowledge on how to
cultivate, preserve and feed Moringa oleifera to dairy cows in order to obtain
more high quality feeds and higher milk yields, mainly during the dry season
under dry tropical conditions in Nicaragua.

Specific objectives of the studies were:

To determine the effect of two plant densities and four levels of
nitrogen fertilisation on the biomass production and chemical
composition of leaves, petioles and stems of Moringa oleifera under
field conditions.
To appraise how Moringa leaf meal (MLM) compares to commercial
concentrate constituents with regard to milk yield, milk
composition and ration digestibility.
To evaluate the effect on fermentation characteristics when Moringa is
introduced for ensiling in various mixtures with at least one of the
components of Elephant grass, sugar cane or sugar cane molasses.
To assess the effect of inclusion of Moringa on aerobic stability of
silages in terms of CO
2
production and time to spoilage.
To investigate the effect on milk yield, milk composition and
digestibility of feeding fresh or ensiled Moringa compared with
feeding a conventional diet of Elephant grass plus commercial
concentrate.
To measure the effect of feeding Moringa, fresh, ensiled or as leaf
meal, on the organoleptic characteristics of milk produced by the
cows on these diets.
16


17
3 Summary of Materials and Methods
3.1 Treatments
An agronomy experiment was performed (Paper I) with two different
planting densities (D
1
= 100,000 and D
2
= 167,000 plant ha
-1
) and four levels
of nitrogen (N) fertilisation as follows: N
1
=0, N
2
= 261, N
3
= 521 and N
4
=782
kg N ha
-1
year
-1
. In Paper II a basal diet of Pennisetum purpureum was offered
to the cows in the study and concentrate containing 20% soybean meal
(SBM) as protein source was compared with a concentrate where SBM was
replaced with the same amount of MLM. In a third diet, commercially
available components were used to compose an Iso concentrate with the
same energy and protein content as the concentrate containing MLM.
To investigate the ensilability of Moringa, 14 different treatments were
tested in Paper III. These were four different combinations of Moringa and
Elephant grass with 10 or 50 g kg
-1
fresh matter (FM) molasses, two
combinations of Moringa and sugar cane, two treatments based on Moringa
with 10 or 50 g kg
-1
FM molasses, one treatment with a combination in
equal parts of Moringa, Elephant grass and sugar cane, two combinations of
Elephant grass and sugar cane, two treatments with Elephant grass with 10
or 50 g kg
-1
FM molasses and one treatment based only on sugar cane.
In Paper IV, Moringa foliage, either fresh or ensiled, was compared with
a conventional diet for dairy cows. A feed ration was planned to fulfil DM
and ME requirements (NRC, 1988) for the control treatment in which 60%
of the expected DM intake was given as roughage in the form of Elephant
grass and the remaining 40% was given as commercial concentrate. Moringa
treatments (fresh or ensiled) were planned to be isocaloric with the control
treatment. As a promoter of palatability, 1 kg molasses was added to the
Moringa treatments.

18
3.2 Locations
The experiments described in Papers I, III and IV were performed at the
farm of the National University of Agriculture (UNA) in Managua,
Nicaragua, geographically located at 120815N; 860936W. The average
annual rainfall is 1440 mm, with a marked dry season (November-May).
The experiment described in Paper II was carried out at the Santa Ana farm
in Masaya, Nicaragua, located at 132916.5N; 605510W.
All experiments were conducted under dry tropical forest conditions.
The soil where Moringa oleifera was cultivated in Paper I belongs to the
taxonomical order of Andosols (FAO, 1988) with clay loam textural class
(USDA, 1995).
3.3 Planting Moringa
3.3.1 Soil preparation and sowing
For the agronomy experiment (Paper I), soil preparation consisted of
conventional tillage using a tractor and mechanical tools to clean the land of
plant debris, followed by disc ploughing, disc cultivation and two
harrowings. Untreated seeds of Moringa were used for propagation. In June
2007, seeds were sown in 2 cm deep holes at the study site (2 seeds per
hole). After 2 months of growth, the stand was thinned and only one
healthy plant was kept. Irrigation was not applied. Weeds were controlled
manually 30 days after germination and every second month throughout the
experiment.
3.3.2 Fertilisation
In this long-term experiment, June and October were set as the fertilisation
occasion. To establish suitable experimental levels of N-fertilisation, data
from a production experiment performed in Nicaragua with Moringa oleifera
(Reyes-Snchez et al., 2006a) were used to estimate expected dry matter
yields and chemical composition of Moringa foliage. The total crude protein
yield on a DM basis reported by Reyes-Snchez et al. (2006a) was divided by
6.25 to get the amount of nitrogen required by the crop, and in the present
experiment this was set as N
3
. Levels 50% below and above this level were
set as N
2
and N
4
, A control treatment with no nitrogen fertilisation was set as
N
1
. The levels of N fertilisation used in this study were therefore N
1
=0, N
2
=
261, N
3
= 521 and N
4
=782 kg N ha
-1
year
-1
. In cuts corresponding to October
and June of each experimental year, N fertiliser was applied 2 weeks after
pruning and was directly incorporated into the soil by manual hoeing.
19
The aim of the fertilisation strategy was to allow the effect of different N
levels without risking a shortage in phosphorus (P) and potassium (K) that
could influence the results. Therefore, for calculation of P and K
requirements the data from Reyes-Snchez et al. (2006a) were also used as
reference for estimating DM yields and chemical composition of Moringa
foliage. Based on the P and K contents in DM, the total amount of those
minerals was calculated and set as 100% of their requirements, but 150% of
the total K and P required was applied at sowing in the entire experimental
area to ensure the requirements would be met, which corresponded to 44
kg of P ha
-1
and 731 kg K ha
-1
, respectively. Urea (46-0-0), triple super
phosphate (0-45-0) and muriate of potash (0-0-60) were used as source of N,
P and K, respectively.
3.4 Feeds
3.4.1 Moringa leaf meal (MLM)
The branches with leaves and soft twigs used for production of MLM for the
experiment presented in Paper II were collected from Moringa trees in an
experimental area by cutting every 45 days. The harvested material was sun-
dried for 24 h before the partially dried leaves were removed by threshing
and then sun-dried again for approximately 48 h on black plastic sheets. The
dried leaves were finely ground in a hammer mill, packed in sacks and
stored in a well-ventilated storeroom.
3.4.2 Concentrates
The experimental concentrate mixtures used in experiments with MLM to
dairy cows (Paper II) and feeding either fresh or ensiled Moringa to dairy
cows (Paper IV) were produced at the Feed Concentrate Plant of the UNA.
All concentrate ingredients were purchased on the local market. Sugar cane
molasses was purchased from an agricultural feed supplier.
3.4.3 Ensiling
Non-irrigated and unfertilised Moringa foliage, Elephant grass (Pennisetum
purpureum) and sugar cane (Saccharum officinarum) were harvested from the
experimental field of UNA and used together with molasses for silage making
in Paper III. All forages were hand-cut by machete 45 days after pruning.
Sugar cane was cut 9 months after the previous cut. All leaves and roots of
the sugar cane were removed from the stems and only the stems were
20
ensiled. After cutting, forages were chopped into pieces of approximately 2
cm in length.
Once the forages were chopped, the 14 treatments with different
proportions of Moringa, Elephant grass and sugar cane were prepared for
ensiling. In treatments without sugar cane, molasses without any water
dilution was added at rates of 10 or 50 g kg
-1
FM. Mixtures for each
treatment were prepared and from these batches fresh material was taken to
fill on average 1564 g FM in glass jars with a nominal volume of 1800 ml.
The fresh material was pressed into the jar to remove as much air as possible.
The glass jars (micro-silos) were fitted with water-locks on their lids to let
fermentation gases escape.
Moringa silage used in Paper IV was prepared 120 days before the
experiment by cutting a field of Moringa 45 days after a previous harvest.
The material was chopped and molasses was added at a rate of 5% according
to fresh weight. The material was then put into 55x97 cm polyethylene bags
and compressed by hand. The bags were sealed to serve as silos. In total 180
bags, weighing approximately 45 kg each, were stored indoors on a concrete
floor until opening.
3.4.4 Fresh forages
Fresh forages such as Moringa and Elephant grass were used as part of the
experimental diets in Papers II and IV. For the basal diet of cows fed locally
produced concentrates (Paper II), Elephant grass cv. Taiwan was used.
However, P. purpureum cv. CT115 was used as part of the control treatment
in Paper IV.
The fields with Elephant grass and Moringa were already regularly
harvested in a harvesting system before the start of the experiments in order
to enable approx. 45-day harvest intervals throughout the experiments. Each
day, Elephant grass and Moringa from a new plot were harvested, chopped
mechanically into 2 cm lengths and offered fresh to the cows.
3.5 Animal Management
Six dairy cows were used in Papers II and IV, each in their second or third
lactation and weighing on average about 450 kg. The cows were in their
fourth week of lactation at the start of the experiments. All cows were
treated according to EC Directive 86/609/EEC for animal experiments.
The animals were weighed at the beginning of the trials and were kept
loose in individual, well-ventilated stalls with a concrete floor. Before the
start of the trials, all animals were injected with Vitamin A (625,000 IU) and
21
Vitamin D
3
(125,000 IU), treated against external and internal parasites and
vaccinated against anthrax. Water and minerals were supplied ad libitum.
Except during the weeks when faecal samples were taken, the cows were
allowed to exercise daily in a common area while the individual stalls were
being cleaned.
Digestibility studies were performed in the feeding experiments (Papers II
and IV). During the last week of each period, rubber mats were placed in
the stalls while all of the faeces from each cow were manually collected.
During the faeces collection period, cows were monitored 24 h per day to
ensure total collection of faeces.
3.6 Sampling
3.6.1 Biomass
At the start of the agronomy experiment (Paper I) in October 2007, the
whole plantation was uniformly cut at a height of 30 cm above the ground
and all foliage was removed but not weighed. The regrowth was harvested
throughout the two subsequent years, starting from mid-November 2007.
Every 45 days, the regrowth was harvested at 40 cm above the ground using
a machete. The fresh biomass from each plot was weighed and recorded to
estimate fresh matter yield. The material obtained from each plot was
separated into two fractions: a fine fraction, which included leaves, petioles
and soft stems 5 mm in diameter or smaller, and a coarse fraction, which
included stems larger than 5 mm in diameter. The weight of each fraction
was recorded and samples of the fine and coarse fraction were taken for
subsequent chemical analysis.
Average height of the plants was estimated by measuring the heights of
five randomly selected plants in each sub-plot of each treatment. The
measurements were made between the plant base (ground) and the highest
tip of the leaves. Mortality was calculated as percentage of plants that died at
the end of the year, divided by the number of live plants at the beginning of
the corresponding year, for each sub-plot.
3.6.2 Feed and faeces
In feeding experiments (Papers II and IV), the amounts of feed offered and
occasional refusals were weighed daily and sampling of feed was performed
as follows. One kilogram of offered roughage per cow and day was collected
and immediately frozen at -18 C. After every period the frozen samples of
22
offered feed for each cow were thawed and pooled into one sample per
period for chemical analysis.
The occasional refused feed was individually sampled and frozen
immediately at -18 C. At the end of each experimental period these
individual samples of refusals were thawed and sent to a laboratory for
chemical analysis. From each of the concentrates used in Papers II and IV, 1
kg from every delivered batch was collected for analysis.
Samples of ensiled Moringa (Paper III) were obtained when micro-silos
were opened 120 days after being sealed. The contents from each micro-silo
were transferred into a plastic bag, thoroughly mixed and then 800 g samples
were taken for analysis. An identical procedure was performed with samples
taken after the aerobic stability study.
During digestibility studies in the feeding experiments (Papers II and IV),
whenever a cow adopted the defecation position, a shovel was put under her
tail to collect the faeces, thereby minimising urine and dirt contamination.
The faeces from each cow were put into a large container and covered with
a lid to avoid evaporation. Once daily, the faeces from each container were
weighed and thoroughly mixed. Five percent were taken as a sub-sample
and frozen before the containers were emptied. When the collection was
completed, the sub-samples from each cow were thawed and mixed
together. Approximately 300 g of this mixture were then taken as a faecal
sample for each animal and experimental period to be used in the chemical
analysis.
3.6.3 Milk
The cows in feeding experiments (Papers II and IV) were hand-milked
twice daily at approximately 12-h intervals. At each milking, the yield was
weighed and recorded and 100-ml samples were taken in sterile vials. The
milk samples were immediately refrigerated at 4 C and then pooled into
one sample per cow and data collection period for analysis of fat, total solids,
CP and casein. The same milk sampling procedure was repeated during the
last three days of each experimental period and samples were submitted to
the laboratory for organoleptic testing on the day after collection.
3.7 Chemical Analysis
In all experiments, the analyses of DM, CP and ash in feed offered, refusals,
faeces and silages were performed using standard AOAC methods. Neutral
detergent fibre (NDF), acid detergent fibre (ADF) and lignin were determined
by the methods of Van Soest et al. (1991) using sodium sulphite. In the
23
ensiling Moringa experiment (Paper III), pH and concentration of lactic
acid, propionic and acetic acid were determined by High Performance
Liquid Chromatography and Clostridium perfringens, lactic acid bacteria,
aerobic enterobacteria and fungal growth were determined according to
APHA (2001). Water soluble carbohydrates in Paper III were determined as
described by AOAC (2000).
The CO
2
concentration in potassium hydroxide solution and pH in silages
from Paper III were analysed according to Ashbell et al. (1990). Nitrogen
content in milk was determined using the Kjeldahl method (AOAC, 1984)
and milk protein using the Babcock method (Pereira, 1988), while total
solids and casein were analysed according to AOAC procedures (Papers II and
IV).
Organic matter digestibility and the metabolisable energy (ME) of
roughages were determined by in vitro incubation in rumen liquid. The ME
was then estimated using an equation presented by Lindgren (1979). The
Weende analysis was used to determine the ME content of the concentrate,
as described by McDonald et al. (1988) (Papers II and IV).
3.8 Organoleptic Characteristics
The organoleptic evaluation of milk in feeding trials (Papers II and IV) was
performed by an experienced panel. A triangle difference test (Witting de
Penna, 1981) was applied using a milk sample with normal sensory
characteristics (flavour, aroma, colour and appearance) as standard.
Table 1. Descriptive terms and scores used in test of organoleptic characteristics of milk (adapted and
abbreviated from Witting de Penna, 1981). Score for aroma and flavour 1-5, and for colour and
appearance 1-3
Score Aroma Flavour Colour Appearance
5 Undoubted
characteristic
Undoubted
characteristic

4 Normal Normal
3 Subtly lack of freshness Subtly impure Yellowish white Homogeneous when
shaken
2 Subtly grassy Subtly grassy Markedly white Presence of small
lumps after shaking
1 Markedly impure Markedly
grassy
Abnormal
coloration
Presence of large lumps
after shaking

Twenty judges were asked to rate the sensory characteristics of milk using
the score sheet presented in Table 1, where 5 was the maximum score for
24
flavour and aroma and 3 was the maximum for colour and appearance. The
total score of each quality trait was the sum of scores of all judges, while
classification was the set of data that occurred most often.
3.9 Experimental Design and Statistical Analysis
The agronomy experiment (Paper I) had a split-plot design with four
randomised complete blocks. The two planting densities were the main plot
factor and the four levels of N fertilisation were the sub-plot factor. In total,
32 plots were established. A repeated measures analysis of variance in the
Mixed procedure of SAS (SAS, 2004) was conducted to determine the effect
of plant density and level of fertilisation on the variables measured in a series
of harvests over two years (Paper I). Tukeys pair-wise comparison was used.
Observations from the same sub-plot were assumed to be correlated with a
first order autoregressive structure. Degrees of freedom were estimated using
the Kenward and Rogers method. The interactions between planting
density-year, planting density-fertilisation level, planting density-cut and the
higher order interactions were tested but then removed from the model due
to lack of significance (P>0.10).
A changeover 3x3 Latin square, as described by Patterson and Lucas
(1962), replicated in two orthogonal Latin squares was used in feeding
experiments (Papers II and IV). Each experimental period consisted of 2
weeks for treatment adaptation and 2 weeks of data collection. Data from
Papers II and IV were analysed using the GLM procedure in SAS version 9.12
(SAS, 2004), while Tukeys pair-wise comparison procedure was used
whenever the overall F-test of treatment means showed a significant result.
Carry-over effects from previous periods as well the interaction between
periods and treatments were tested initially, as described by Patterson and
Lucas (1962), but were excluded from the final model because of lack of
significance (P>0.10).
A completely randomised experimental design with 14 treatments and
three replicates of each treatment was used in the experiment on the
ensilability of Moringa (Paper III). The GLM procedure in SAS (2004),
including the options MEANS, PDIFF and ESTIMATE, was used. First and
second-order mixture experiment models as described by Atkinson and
Donev (1992) were used for the data in Paper III. Goodness-of-fit tests were
performed using the differences of the sums of squares of the one-way
model and of the mixture experiment model as the numerator in an F-test.
In cases where a second-order model was applied, the product effects were
successively tested and removed until only significant products remained. In
25
parallel with significance testing of mixture experiment models, descriptive
correlation coefficients of the predicted values and the treatment means were
calculated. Variables not showing sufficient goodness-of-fit with the mixture
experiment models were analysed by the one-way model only.

26


27
4 Summary of the Results
4.1 Moringa as a crop (Paper I)
The agronomy experiment was conducted as a field experiment from June
2007 to October 2009 in an area which was previously fallow land. Year
one consisted of the dry season of 2007 and the rainy season of 2008, while
year two was the dry season of 2008 and the rainy season of 2009. There
were great variations in precipitation between years, with rainfall of 1411
and 1439 mm in 2007 and 2008, but with unusually low rainfall during May
and June of 2008 and very dry conditions in 2009 (796 mm). There were
no significant differences (P<0.05) between planting densities with regard to
biomass production. The density 167,000 plant ha
-1
(D
2
) produced the
highest total dry matter yield (TDMY) and fine fraction dry matter yield
(FFDM), with 21.2 and 19.2 ton ha
-1
respectively, while 11.6 and 11.0 ton
ha
-1
were obtained for the same variables at the density 100,000 plant ha
-1

(D
1
). Growth rate (GR) in D
2
reached a maximum value of 0.06 compared
with 0.03 ton ha
-1
day
-1
in D
1
. Plant height (H) was on average 119 cm and
no differences were found between planting densities.
There were no significant differences between N
3
and N
4
in any year with
regard to TDMY. On average, those levels produced 22.5 and 27.7 ton ha
-1

year
-1
during 2007-2008 and 2008-2009, respectively. Both were
significantly higher (P<0.001) than levels N
1
and N
2
. During 2007-2008, the
TDMY for levels N
1
and N
2
were 8.9 and 13.9 ton ha
-1
year
-1
, while the
corresponding values for the same treatments in 2008-2009 were 4.7 and
10.0 ton ha
-1
year
-1
. N
2
showed no significant differences between years,
while a significant (P<0.05) reduction in TDMY of 47% was observed in N
1

during 2008-2009 compared with the previous year. There were significant
interactions between year and fertilisation level in terms of FFDM and GR.
28
Levels N
1
and N
2
gave higher yields of FFDM and higher GR in 2007-2008
than in the following year, while N
3
and N
4
showed an increase in 2008-
2009 compared with the previous year for these variables. No interactions
were found between year and fertilisation level for plant height. Plants were
consistently taller in 2007-2008 than in the following year.
There were significant interactions between cuts and years with regard to
biomass production. A maximum yield of well over 5 ton ha
-1
TDMY was
observed in Cut 6 in May of both years. Cut 6 also gave a significantly
higher (P<0.05) FFDM in 2008-2009 than in 2007-2008. In addition, a
significant interaction (P<0.01) was found between year and cut for GR,
which was at least 25% lower in Cuts 4 and 6 in 2008-2009 compared with
2007-2008. There was a significant (P<0.05) interaction between year and
cut for H, which was higher during the first three cuts in 2008-2009
compared with 2007-2008.
There were significant interactions between fertilisation level and cut
with regard to FFDM, GR and H. The levels N
1
and N
2
consistently differed
significantly (P<0.05) from the higher fertilisation levels on all cutting
occasions. At Cut 6, levels N
3
and N
4
each produced just over 5 ton ha
-1

FFDM, while 3 and 2 ton ha
-1
were obtained on average for the same fraction
with levels N
1
and N
2
, respectively. The same pattern was observed in GR,
while plant height (H) from different fertilisation levels followed more or less
the same pattern.
No significant differences were found between planting densities with
regard to the chemical composition of plant fractions. On average, the
content of DM, CP, NDF, ADF and lignin in the fine fraction was 115 g kg
-1
and 276, 352, 241 and 82 g kg
-1
DM respectively, while 203 g kg
-1
and 72,
670, 683 and 109 g kg
-1
DM were found for the same variables in the coarse
fraction. No significant differences were found among fertilisation levels
with regard to CP content in the fine fraction. However, a significantly
(P<0.05) low value (71.7 g kg
-1
DM) of CP content was obtained in the
coarse fraction in level N
1
.
4.2 Moringa leaf meal in concentrates (Paper II)
In an experiment where MLM was used as an alternative protein source to
locally produced concentrates, no significant differences were found
between the two isocaloric and isoproteinic treatments with regard to
digestibility or milk yield. The SBM treatment displayed higher digestibility
in CP and produced significantly (P<0.05) more milk (13.2 kg cow
-1
day
-1
)
29
compared with the MLM treatment (12.3 kg cow
-1
day
-1
), in line with the
higher allocation of CP, ME and starch.
There were no differences in milk composition. The milk contained 34.9
g fat, 34.5 g protein, 126.1 g total solids and 27.4 g casein per kg. No
differences in colour, aroma and flavour were found between treatments and
all of them were classified as normal.
4.3 Moringa as silage (Paper III)
The silages prepared with different combinations of Moringa, Elephant grass
and sugar cane fell within a narrow DM range of 210-269 g kg
-1
, where
Elephant grass treatments had the highest DM content regardless of their
content of molasses. Furthermore, CP concentration increased markedly
with increasing proportion of Moringa, with the highest CP concentration
being found at the highest proportion of Moringa (treatment with 99% of
Moringa and 1% molasses). The concentration of NDF in silage decreased
when the proportion of Moringa increased.
The presence of Moringa in treatments decreased pH values by 0.8
(P<0.001). Correspondingly, the presence of Elephant grass increased pH
values by 0.7 (P<0.001). No such effect was seen with the presence of sugar
cane. Both Moringa and Elephant grass, as well as the proportion of
molasses, affected the lactic acid concentration, by 16 g kg
-1
DM (P<0.001), -
21 g kg
-1
DM (P<0.001) and -12 g kg
-1
DM (P<0.05) respectively. The
presence of sugar cane decreased acetic acid concentration (P<0.05).
A tendency for higher Clostridia numbers was observed in treatments
with sugar cane rather than with molasses. Intermediate LAB numbers were
found in treatments with Moringa and no differences were found among
treatments with regard to fungal growth. All silages had Enterobacteria
numbers below the detection limit of 0.5 most probable number (MPN) g
-1
.
The presence of Moringa caused a significant (P<0.001) increase in CO
2

production, decreased time to spoilage by 67 h (22%) (P<0.05) compared
with the Elephant grass silages and increased pH after spoilage by 1.55
(P<0.001).
4.4 Moringa as roughage (Paper IV).
When Moringa, either fresh or ensiled, was tested as the sole roughage in
the diet of dairy cows, intake of DM was approximately as planned. There
were no refusal in the control treatment and feed refusals in the Moringa
treatments were minor, representing only 0.97% of the daily amount
30
offered. No differences between Moringa silage (MS) and Moringa fresh
(MF) treatments were found in terms of apparent digestibility coefficient,
with the exception of DM, where MS gave the highest value of 0.76
compared with 0.64 and 0.69 in the control diet and MF, respectively.
Moringa silage treatment had significantly higher (P<0.05) values than the
Elephant grass+concentrate treatment with regard to CP, OM, NDF and ADF
digestibility, with values of 0.83, 0.77, 0.66 and 0.61, respectively.
There were no significant differences between treatments with regard to
milk yield and energy corrected milk (ECM), which averaged 13.7 and 12.9
kg cow
-1
day
-1
, respectively. Furthermore, there were no significant
differences in milk composition between treatments and on average the milk
contained 35.1 g fat, 34.5 g protein, 123 g total solids and 27.3 g casein per
kg.
The colour and appearance of milk from all treatments were classified as
normal and no differences in these traits were found between treatments.
However, milk from the MF treatment was classified as subtly grassy and was
significantly different (P<0.001) to the other treatments, which were
classified as normal with regard to flavour and aroma.


31
5 General Discussion
5.1 Biomass production
Significant levels of edible biomass production and tolerance to pruning are
two important factors determining the suitability of fodder trees or shrubs as
forage species (Benavidez, 1996). Moringa has proven to be tolerant to
pruning under different cutting frequencies (Reyes-Snchez et al., 2006a;
Foidl et al., 2001). In terms of the biomass production factor, Paper I
evaluated the effect of planting density and level of nitrogen fertilisation on
biomass yield. The total biomass produced was divided into two categories
(Figure 1), a fine fraction with leaves and stems less than or equal to 5 mm
in diameter and considered easily edible for dairy cows, and a coarse fraction
with stems of diameter over 5 mm and usually not considered edible.
However, under Tropical Dry Forest conditions this coarse fraction has been
observed to be consumed by steers and goats and could warrant further
investigation.


Figure 1. Moringa oleifera foliage divided into (a) fine fraction and (b) coarse fraction. The ruler
in the picture is 30 cm long.
a b
32
A positive relationship between planting density and biomass yield in
tropical tree legumes has been reported (Ella et al., 1989), although Reyes-
Snchez et al. (2006a) and Manh et al. (2005) observed no effect of planting
density on biomass yield of Moringa. Paper I showed that of the two
densities studied, the higher density (167, 000 plant ha
-1
) gave a higher yield
of Moringa. The TDMY obtained at that density (21.2 ton ha
-1
) was higher
than the 18.9 ton ha
-1
reported by Reyes-Sanchez et al. (2006a) at a planting
density of 750,000 plants ha
-1
.
The main reason for the difference is that at the high densities (250,000
to 750,000 plant ha
-1
) studied by Reyes-Snchez et al. (2006a), the
competition for nutrients and sunlight among plants was observed to be
high. At lower planting densities such as those in the present experiment
(100,000 and 167,000 plants ha
-1
), plants do not need to compete as much
and therefore produce higher yield.
Fertilisation is a key point in cut-and-carry systems, where huge amounts
of nutrients are removed from the areas where the crop is harvested. In a
perennial crop such as Moringa this leads to a reduction in biomass yields
over time, especially if high planting densities are used. The TDMY in 2007-
2008 was 8.9, 13.9 and 22.5 ton ha
-1
for N
1
, N
2
and the average of N
3
and N
4
respectively, while in 2008-2009 the TDMY was 4.7, 10 and 27.7 ton ha
-1
for
the corresponding treatments. As can be seen, dry matter production
increased yearly at the two higher levels of N (N
3
=521 and N
4
=782 kg N ha
-1

year
-1
) but declined with N
1
=0 and N
2
=261 kg N ha
-1
year
-1
. The response of
plant growth to N is widely recognised in conditions where N is a limiting
factor (Salmern-Miranda et al., 2007).
There were important external factors besides the experimental
conditions affecting the biomass production in Paper I. Biomass production
followed the rainfall pattern and the sum of DM yield from the rainy season
(May-October) was twice that harvested during the dry season (November-
April). It is important to highlight that while the interaction between years
and cuts was not very clear during the dry season (Cuts 1 to 4), the drier
second year (2008-2009) produced larger yields with regard to FFDM
compared with the first year (2007-2008). Plants tend to grow thinner in
dry conditions compared with wet, where plants not only grow taller but
also thicker, with a higher coarse fraction (Mommer et al., 2006). In general,
higher levels of fertilisation generated higher yields in the rainy season.
Furthermore, the physiological response of the plant itself to repeated
cutting should be considered a factor that affects the proportion of fine or
coarse fraction, as well the total dry matter yield. When Moringa was newly
established it developed only one trunk, which was fairly thick. However, as
33
soon as pruning started the regrowth became thinner and more branched.
This can also be seen in data reported by Reyes-Snchez et al. (2006a),
where the coarse fraction was reduced by 60% in the second year compared
with the first.
5.2 Characterisation of Moringa as a feedstuff
In tropical countries, including Nicaragua, forage quality is often too low to
meet the nutritional requirement of animals. Furthermore, supplementation
with conventional concentrates is generally too costly and the levels of
concentrate feeding are therefore low. New low-cost alternatives to
commercial concentrates are needed and Moringa has been shown to be one
possible option. However, the first critical step in its general use in livestock
diets is precise and reliable knowledge of its chemical composition,
digestibility and nutritional value. Other practical issues in connection with
the use of Moringa as a feedstuff are the labour requirement and how well it
can be conserved. Such information is particularly vital in the current
context, where farmers are trying to achieve more sustainable production
throughout the year.
5.2.1 Chemical composition of Moringa foliage
Moringa foliage can be used either fresh or dry in animal diets depending on
the species and production aim. Fresh foliage was used in the diet of dairy
cows in Paper IV and the chemical composition of the fine and coarse
fractions of Moringa foliage was determined in Paper I.
The DM in 45-day-old fresh foliage in this study was in the range 110.4-
203.8 g kg
-1
. A variation in DM content is also reported in the literature, e.g.
Reyes-Snchez et al. (2006a and 2006 b) found a range of 164-228 g kg
-1
.
This variation can partly be explained by the cutting frequency of the
material used, as can be seen in Reyes-Snchez et al. (2006a) where the DM
content of fresh foliage increased from 164 to 228 g kg
-1
in plants aged 45
and 75 days, respectively. The thickness of the material analysed can also be
an important factor influencing the DM content. In this study fresh foliage
was divided into two fractions, a fine fraction which included leaves,
petioles and stems with diameter less than or equal to 5 mm, and a coarse
fraction with diameter above 5 mm. When the fine fraction was analysed,
the DM content was consistently between 110 and 120 g kg
-1
(Paper I),
while it was about 200 g kg
-1
for the coarse fraction. A DM content of up to
460 g kg
-1
has been reported for the coarse fraction with branches, stems,
petioles and leaves analysed together (Aregheore, 2002).
34
Different authors have cited Moringa as a high protein content fodder,
with CP concentrations up to 290 g kg
-1
DM (Aregheore, 2002; Foidl et al.,
2001). In this study, the CP concentration in Moringa foliage fell within a
narrow range of 241 and 277 g kg
-1
DM in the fine fraction (Papers I and
III). However, the CP concentration decreased to 70 g kg
-1
DM when only
the coarse fraction was analysed (Paper I). The relationship between plant
fraction and CP concentration has been reported previously. Makkar and
Becker (1997) reported 264 g kg
-1
DM in leaves, 72 g kg
-1
DM in twigs and
62 g kg
-1
DM in stems. That CP value for leaves is similar to the CP content
in the MLM (292 g kg
-1
DM) used in Paper II.
The cell wall content in foliage is highly variable and influenced by
certain factors, such as species, phenological stage at harvest and preservation
method. The NDF content of Moringa foliage was between 510 and 521 g
kg
-1
DM (Papers III and IV) but when fraction differentiation was performed
in the production trial, 348 g kg
-1
DM was obtained for the fine fraction and
683 g kg
-1
DM for the coarse fraction (Paper I). This is consistent with
reports from other tropical fodder trees such as Morus alba, Gliricidia sepium,
Guazuma ulmifolia and Sesbania grandiflora, which have an NDF content in
leaves of between 281 and 570 g kg
-1
DM and in stems of between 638 and
720 g kg
-1
DM (Sultan et al., 2008; Solorio-Snchez et al., 2000).
The in vitro dry matter digestibility (IVDMD) of Moringa foliage (0.69) did
not differ greatly from published values in other tropical multipurpose trees
(Sultan et al., 2008; El hassan et al., 2000; Solorio-Snchez et al., 2000) and
in other reports about Moringa (Reyes-Snchez et al., 2006a). This
digestibility is at the level of alfalfa hay or maize silage (Holden, 1999). This
information, together with the CP and fibre content, is particularly
interesting because Moringa fodder is intended to be used as a protein
supplement for low-quality tropical fodders or even as the only source of
roughage for dairy cows.
In Paper I, there were no obvious effects of N fertilisation level on CP,
ADF and ash content of the fine fraction. The effect of N fertilisation on
chemical composition of foliage seems to vary among species of fodder trees,
but no other reports on this in Moringa have been found for comparison.
When the shrub Manihot sculenta was studied, a significant effect of
fertilisation on CP, NDF, ADF and ash content was reported (Phengvichith et
al., 2006), but with the tree Morus spp., Rodriguez et al. (1994) reported no
such effect. Different responses to N fertilisation can also be seen among tree
species. Hartley et al. (1995) found significant differences in N content on
foliage of Picea sitchensis depending on fertilisation, but no effect on chemical
composition of Calluna vulgaris foliage in the same experiment.
35
Moringa can be also dried in different ways. In Paper II it was sun-dried
on plastic sheets in open sun in order to obtain Moringa leaf meal (MLM).
This type of meal has been tested as a feed for different species in recent
decades, as shown in Table 2. Based on the information in Table 2, some
conclusions can be drawn. The first is that the nutritional effects obtained
when feeding Moringa to different species are mixed; and the second is that
experiments where MLM has been used as a feed for dairy cows are still few.
Table 2. Use of Moringa leaf meal (MLM) and its nutritional effects in different species.
Inclusion level Species Nutritional effect Source
10 % of total
dietary protein
Nile tilapia Same growth rate as
commercial concentrate
Richter et al., 2003
Substitution of
30 % of
fishmeal
Nile tilapia Same growth rate as
commercial concentrate
Afuang et al., 2003
4.3 to 42.5% of
total diet
Nile tilapia Reduced growth
compared with a
commercial concentrate
Dongmeza et al., 2006
45.2 % of total
diet
Abalone Improvement in yield Reyes and Fermin, 2003
Up to 20% of
total diet
Rabbits Leaner carcass Nuhu, 2010
10 % of total
diet
Laying hens Unclear effects on egg
weight
Kakengi et al., 2007
Up to 5% of
total diet
Broiler
chickens
Reduction in
performance in terms of
weight gain, feed
conversion ratio and final
body weight at 8 weeks
when above 5% of total
diet
Olugbemi et al., 2010
20% of total
diet
Growing sheep 20% improvement in
growth rate but poorer
feed conversion
Murro et al., 2003
1.65 kg DM Cross-breed
cows
Same milk yield as cows
fed 1.23 kg DM
cottonseed meal
Sarwatt et al., 2004

Under the conditions in which MLM was produced in this study (Paper
II), the chemical composition obtained was 922 g kg
-1
DM and 292 g CP,
161 g NDF, 151 g ADF, 68 g lignin, 94 g ash and 10.9 MJ EM kg
-1
DM. A
broad variation in nutrient content has been reported for MLM, with CP
concentration varying from 120 to 349 g kg
-1
DM (Madalla, 2008; Murro et
36
al., 2003); NDF from 159 to 320 g kg
-1
DM (Sarwatt et al., 2004; Richter et
al., 2003); and ADF from 44 to 320 g kg
-1
DM (Sarwatt et al., 2004; Afuang et
al., 2003). Ash content has been reported to range from 71 to 194 g kg
-1
DM
(Nuhu, 2010; Sarwatt et al., 2004).
There are many different reasons for those variations. However, the main
reason that chemical composition of MLM can differ considerably is that
sometimes a certain amount of smaller branches and twigs is included along
with the leaves in the leaf meal and the proportion of this non-leaf fraction
has a large impact on the chemical composition. Fujihara et al. (2005)
reported a decrease of 22% in CP concentration when soft twigs were
included along with leaves in leaf meal compared with leaves alone. Other
authors reported 254 g kg
-1
DM when only leaves were used in leaf meal and
120 g kg
-1
DM when leaves and branches were used (Afuang et al., 2003;
Murro et al., 2003). The CP content of soft twigs alone is lower but this
fraction can be used for animals with lower nutrient requirements such as
dry cows, which readily consume this feed.
Reyes and Fermin (2003) have also suggested differences in agro-climatic
conditions and different age of trees as a source of variation in the chemical
composition when material is collected from uncultivated trees. Pok et al.
(2005) performed an experiment where new leaves (recently open at the top
of braches) and old (near to yellow colour at the bottom of branches) were
compared with regard to protein concentration and N digestibility. No great
variation was found in crude protein concentration between new and old
leaves, which contained 32.7 and 30 g kg
-1
DM, respectively. However, a
46% decrease in in vitro N digestibility (IVND) was reported. Based on basic
plant physiology, a higher concentration of CP and a lower content of fibre
can be expected in younger tissues than in older.
The drying method can also be considered a factor in variations in
chemical composition. There are different procedures to obtain MLM;
freeze-drying has mainly been used in aquaculture trials (Dongmeza et al.,
2006; Afuang et al., 2003; Reyes and Fermin, 2003; Richter et al., 2003),
while air and shade drying have been reported by other authors (Olugbemi
et al., 2010; Kakengi et al., 2007; Murro et al., 2003). The sun-drying
method used in Paper II allows dried leaves to be easily removed from the
coarser fraction, giving a highly digestible product with a high nutrient
content similar to that reported by other authors (Nouala et al., 2006;
Nouala, 2004; Sarwatt et al., 2004). Even though CP concentration seemed
not to be significantly affected by drying method (Olsson and Wilgert, 2007;
Atega et al., 2003), oven and air-drying under a roof has been shown to
increase the cell wall content significantly (Atega et al., 2003).
37
Both the high CP concentration and the high rumen degradability
reported (Fujihara et al., 2005; Soliva et al., 2005; Makkar and Becker, 1997)
suggest that MLM can be used as a supplement for animals mainly in tropical
areas, where basal diets are CP deficient. The rumen degradability of MLM
has been reported to be similar to that of soybean and rapeseed meal and
MLM also seems to promote rumen microbial protein synthesis due to the
substantial contents of readily fermentable N and energy (Soliva et al., 2005).
However, Fujihara et al. (2005) showed that the proportion potentially
degraded in the lower digestive tract was lower than for Leucaena
leucocephala.
5.2.2 Chemical composition of Moringa silage
Even though Moringa is not a legume, it shares characteristics with
leguminous plants that might need to be taken into account in the silage
making process, such as the high CP content in foliage and very low content
of water soluble carbohydrates (WSC). Reported levels of WSC in Moringa
vary from <50 to 110 g kg
-1
DM (Mendieta-Araica et al., 2009; Mustapha
and Babura, 2009) and differences could be mainly attributable to the
different detection method used in those studies. In Paper III, 10 and 50 g
kg
-1
DM of molasses were used in pure Moringa silages as a way to increase
the rapidly fermented carbohydrates.
Moringa ensiled either pure or in mixtures with Elephant grass or sugar
cane substantially increased the CP content of the silages and produced good
silage in general (Paper III). When Moringa was ensiled pure with only 50 g
kg
-1
DM molasses the DM varied from 212 to 267 g kg
-1
, which is within the
adequate range for silages (Buxton et al., 2003). However, the CP
concentration varied from 144 to 226 g kg
-1
DM and the NDF concentration
from 397 to 435 g kg
-1
DM (Papers III and IV). Those differences could be
due to the use of only leaves and soft twigs in Paper IV compared with twigs
and branches in Paper III.
In silages made from tropical grasses, a pH value of 4.2 has been reported
as the maximum to consider silage well-preserved (Crdenas et al., 2003;
McDonald et al., 2002). Weissbach (1996) presented a critical limit for good
quality silage depending on DM content in which pH should be no higher
than 0.0257xDM content (expressed as percentage) + 3.71. Moringa silages
with either 10 or 50 g kg
-1
DM molasses were within the range of good
silages according to these criteria. Moringa silages had higher lactic acid
concentrations (93 to 106 g kg
-1
DM) than silage made from Elephant grass
or sugar cane. The highest concentration of acetic acid was obtained with a
mixture of 0.99 Moringa and 0.01 molasses (31 g kg
-1
DM). Even so,
38
Moringa silages never reached 60 g acetic acid kg
-1
DM, which is considered
the upper recommended level (Crdenas et al., 2003).
Microorganisms are always present in the fermentation process and
depending on their final fermentation products and the way they degrade
nutrients, they can be regarded as beneficial or detrimental. As main genera
lactic acid bacteria (LAB), clostridia, enterobacteria and fungi were studied in
Paper III. Intermediate LAB numbers (3.6-2.5 log CFU g
-1
) were found in
Moringa silages. These microorganisms are necessary to preserve high-
moisture forages as silage. Even though LAB levels of at least 3.9 log CFU g
-1

are desirable for a good fermentation process in temperate grasses, lower
values have been reported as normal in silages from tropical grasses (Pedroso
et al., 2005; Tjandraatmadja et al., 1994). There was a discrepancy between
the lactic acid concentration obtained in the mixture with 95% Moringa and
5% molasses and the LAB count (2.5 log CFU g
-1
), demonstrating that counts
of viable bacteria and lactic acid concentrations do not always coincide.
Clostridia can be regarded as unavoidable in raw material. However, when
found in silages they can lead to undesirable characteristics in the final
product, such as bad smell or increased DM losses. Under the conditions in
which the silages in Paper III were produced Clostridium perfringens was
identified in the silage. Even though the Clostridium levels in Moringa
silages were 3.6-3.8 log CFU g
-1
DM, these values were below the limit of 5
log CFU g
-1
DM reported as a maximum permissible level for good silage
(Lindgren, 1990). Fungal growth was very low and no differences were
found between silages. Enterobacteria numbers were below the detection
limit of log 0.5 MPN g
-1
in all silages.
5.2.3 Labour requirement for silage making and leaf meal production
Nicaraguan farmers at the small and medium scale usually work under
conditions where cash availability is limited and the main resource accessible
is labour, either of the individual or as a family. Therefore the labour
requirement for Moringa silage making and Moringa leaf meal (MLM)
production is important information in helping such farmers plan their
feeding systems. The figures presented here are based on the harvests that
were carried out in Papers I-IV.
At Moringa planting densities of 100,000-167,000 plants ha
-1
and a
harvesting interval of 45 days, the work required to harvest 1 ton of
Moringa foliage is about 19 man-hours. One common way to make silage
under the conditions prevailing on small and medium-scale Nicaraguan
farms is in 55x97 cm polyethylene bags with approximately 45 kg capacity.
The material to be ensiled is put into the bags in layers of approximately 20
39
cm and pressed by hand until the top of the bags is reached, whereupon it is
sealed. Therefore, the labour calculations were based on this technique. To
chop up 1 ton of Moringa foliage using a mechanical chopper requires 1
man-hour. To fill up, compress and seal one plastic bag as a silo using the
technique described above, another 1 man-hour is needed. Therefore to
prepare the approximately 22 plastic bag silos that can be obtained from one
ton of fresh Moringa (about 260 kg DM) requires 42 man-hours.
For MLM production the harvesting labour requirement is the same as
above (19 man-hours). The work involved in spreading out the ton of
Moringa foliage, threshing and final drying on plastic sheets is 6.4 man-
hours. The total amount of work required to produce dry Moringa leaves
from 1 ton of fresh Moringa (about 120 kg DM) is therefore 25.4 man-
hours.
5.2.4 Practical implications of using different feed products from Moringa
Even though Moringa is a good source of protein for dairy cows and can
help farmers overcome the strong effect of dry season feed shortages on milk
yield, there are several practical implications that should be kept in mind
when using feed products from Moringa.
Fresh Moringa has good intake characteristics, but it is necessary to have
an adaptation period to allow cows to get used to the feedstuff. Based on the
experience of the author, this period is never longer than two or three days.
A simple strategy to overwhelm the small distrust of animals facing the feed
can be to offer small amounts of Moringa mixed with other forages and
gradually increase the amount of Moringa and reduce the other forages.
Another option can be to use a small amount of a palatable agent such as
molasses. Once the animals start to eat Moringa, intake does not seem to be
a problem and animals often consume considerable amounts, as seen when
fresh and ensiled Moringa were offered in Paper IV.
Moringa can be used in a cut-and-carry system where the daily forage
requirement can be harvested from the field every morning and offered to
the cows. In contrast to Elephant grass, which needs to be cut and ensiled at
the right stage of development to ensure silage quality (Reyes-Sanchez et al.,
2008), Moringa does not need to be ensiled to ensure high CP content and
high digestibility of the DM. However, it is practical to feed Moringa from a
silo rather than harvesting and transporting the roughage daily. Furthermore,
based on the pattern of biomass yield (Paper I), the fodder production is
abundant in the rainy season and not preserving the foliage implies a
complete loss of surplus production.
40
5.3 Effect of Moringa on milk yield and milk
composition
There are many factors affecting milk yield, genetics and management
among them. However, in tropical areas shortage of feeds in terms of both
quantity and quality is the most important constraint. This is particularly
pronounced during the dry season. In countries like Nicaragua, dry season
milk yield can decrease to 40% of the rainy season milk yield (Fujisaka et al.,
2005).
Moringa has been reported to increase milk yield (Reyes-Snchez et al.,
2006b; Sarwatt et al., 2004). However, it is important to highlight that those
studies were performed with low yielding creole cows (3 to 6 litres milk
cow
-1
day
-1
) and low quality basal diets. Papers II and IV were performed
under medium-scale farm conditions, which can be summarised as follows:
Specialist dairy breeds producing approximately 16 kg of milk d
-1
, two
milkings per day, stall-feeding of planted forage; mainly Elephant grass
(Pennisetum purpureum) as roughage and supplementary feeding with either
molasses, locally available by-products or commercial concentrates at rates of
1 to 10 kg d
-1
(de Leeuw et al., 1998). The main aim in Papers II and IV was
to determine whether Moringa can support the same milk yield as a
control diet representative of the typical diets used in the above-described
milk production system using Elephant grass + commercial concentrate to
cover the nutrient requirements of the cows.
Even though the soybean meal diet in Paper II gave a higher milk yield,
the difference compared with the treatment using Moringa as a source of
protein in concentrate was only 7%. The main reason for differences in milk
yield in Paper II was probably the higher ME and CP intake when cows were
fed soybean meal concentrate. On the other hand, when MLM was
compared with another concentrate with the same nutritional concentration
no differences were found between them, indicating that Moringa as a
protein source has a similar value with regard to milk yield as the
commercially available concentrate in Nicaragua. In this case these
constituents were sorghum, peanut meal and soybean meal.
In Paper IV, where fresh and ensiled Moringa were compared with a
conventional diet with Elephant grass and concentrates, there was no
difference in milk yield between the treatments, although the Moringa
treatments had higher CP and ME intake. This was probably mainly because
no response to extra protein supplementation can be expected when the ME
and CP requirements of cows are met (Broderick, 2003; Oldham, 1984).
Milk composition was not affected by any of the treatments where
Moringa was fed (Papers II and IV). This is consistent with other studies
41
where no relationship between milk protein and percentage of dietary CP
has been observed when the energy concentration in the diet is similar
(Reyes-Snchez et al., 2006b; Schingoethe, 1996; Sutton, 1989).
5.4 Effect of Moringa on organoleptic characteristics
of milk
There is some evidence in the literature that feeding fresh Moringa to dairy
cows can cause off-flavour and aroma (Agrodesierto, 2010). This was fully
confirmed in Paper IV when feeding fresh Moringa. Avoiding feeding fresh
Moringa before the morning milking has been suggested to decrease the
problem, but up to now this has not been confirmed experimentally.
Furthermore, Paper IV showed that when Moringa silage was fed instead
fresh Moringa, no problems in organoleptic characteristics were detected at
all. The possibility of using surplus production during the rainy season by
producing silage, combined with the finding that the milk of silage-fed cows
was characterised by a good flavour and aroma, seems to indicate that there
is potential in the production of Moringa silage for dairy cows.
5.4.1 Flavour
Quoting WHO & FAO (2007), milk is the normal mammary secretion of milking
animals without either addition to it or extraction from it, intended for consumption as
liquid milk or for further processing. However, to be consumed, organoleptic
characteristics such as flavour, aroma, colour and appearance are taken into
account by consumers.
Normal milk has a bland but characteristic milk flavour that is pleasing
and slightly sweet, mainly due to the presence of fat globules, salts and
lactose (Nursten, 1997). Flavour is also claimed to be the most important
attribute for consumer acceptance and preference (Croissant et al., 2007;
Thomas, 1981). Many studies have reported on the relationship between
milk flavour and dietary factor or factors linked to animal management
(Kalac, 2010; Martin et al., 2009; Croissant et al., 2007; Martin et al., 2005).
The transmission mechanism by which flavour substances can be transmitted
to the milk via either the digestive or respiratory route have been described
earlier (Dougherty et al., 1962; Shipe et al., 1962).
Milk is very susceptible to off-flavour, which can originate from multiple
causes. These causes can be divided into two main categories: those
originating from secondary metabolites and those which are transferred to
milk from the environment.
42
A number of compounds are regarded to contribute to the off-flavour in
milk, including metals, terpenes, linolenic acid oxidation products, esters,
glucosinolates, phenolics, phytol derivates and nitrogen heterocycles
(Bendall, 2001; Makkar and Becker, 1997; Walker and Gray, 1970; Shipe et
al., 1962). Some authors have reported the presence of glucosinolates (4-(-
L-rhamnopyranosyloxy)-benzylglucosinolate) in Moringa foliage as the
reason why Moringa gives off-flavour to milk (Bennet et al., 2003; Makkar
and Becker, 1997; Walker and Gray, 1970). However, many authors have
also reported that glucosinolates are susceptible to degradation or are highly
reduced by heat or ensiling (Oerlemans et al., 2006; Panciera et al., 2003;
Vipond et al., 1998; Fales et al., 1987; Nash, 1985). That is presumably the
reason why no off-flavour was found in milk from MLM or Moringa silage
treatments, whereas milk from fresh Moringa was classified as grassy.
5.4.2 Aroma
It is not easy to define milk aroma. It is commonly described as characteristic
but bland, mainly due to the fact that the concentration of aroma
compounds in fresh milk is very low (Bendall, 2001). Some even suggest
that the real fresh milk aroma is a belief rather than a fact (Nursten, 1997).
Due to its blandness, milk can easily get an off-odour.
A very complex combination of different compounds is needed to get
the characteristic milk aroma and more than 70 compounds have been
reported so far (Bendall, 2001). Compounds such as indole and skatole have
been reported to give a faecal smell to milk (Brudzewski et al., 2004).
However, neither of these has been found in Moringa leaves (Bennett et al.,
2003).
The organoleptic characteristics of the milk in Paper IV was analysed
through sensorial analysis, using the method of Witting de Penna (1981).
Using the same methodology as in Paper IV, no effect on milk aroma was
reported by Reyes-Snchez et al. (2006b). This may be due to the small
amount of Moringa fed to the cows (3 kg DM) in that experiment, and the
long time interval between feeding and milking (24 h). In contrast, in the
experiments reported in this thesis, the treatments containing Moringa
constituted 100% of the roughage (Paper IV) or 8% of the total diet (Paper
II) and cows were milked twice a day. The time interval between feeding
Moringa and milking was 1 h and 12 hours after morning and afternoon
milking, respectively.
Milk from cows fed fresh Moringa was classified as subtly grassy and this
can be attributed to the presence of thiocarbamate glycosides in Moringa leaf
tissue (Faizi et al., 1995). Thiocarbamate glycosides have been recognised as
43
odour-active substances (Breme et al., 2007) associated with odours
resembling garlic, onion tops, mushroom and cress-like, to fresh, spearmint-
like in other vegetables from the orders of cruciferous and brassica plants
(Breme et al., 2007; MacGregor, 2000; Walker and Gray, 1970).
The mechanism by which milk can gain a bad smell could either be
through the animal (via digestive or respiratory route) or through the stall
atmosphere. Due to their volatility, thiocarbamates are greatly reduced by
silage making and in the dehydration process used to obtain leaf meal, as has
been reported by many authors (Figueiredo et al., 2007; Oerlemans et al.,
2006; MacGregor, 2000). This explains the absence of a negative effect on
aroma of milk from cows fed Moringa silage or MLM (Papers IV and II).
5.4.3 Colour and appearance
Milk colour is due to a combination of many compounds such as casein,
carotenoids and fat globules and the concentration of those compounds in
milk is related to breed, parity, physiological stage, production level and
sanity state. However, nutrition also plays a very important part, especially
for milk fat content and the carotenoids, which can affect colour. Casein and
fat are regarded as giving a white colour to milk, while carotenoids give a
yellowish shade. The carotenoids are taken up from the blood by the
mammary gland (Martin et al., 2005).
A relationship between more yellow milk and pasture-based diets
compared with silage or concentrate-based diets has already been reported
(Kalac, 2010; Martin et al., 2009; Croissant et al., 2007). However, the diets
used in this study (Papers II and IV) showed no effect on milk colour, even
though fresh forage, silage or a high proportion of concentrate were used.
The reasons for this can be found in four main aspects: a) all the treatment
diets used in Papers II and IV generated the same concentration of casein in
the milk; b) the method used for colour determination was visual and has
been proven not to be as accurate as a spectrocolorimeter (Nozire et al.,
2006a; Nozire et al., 2006b); c) the effect of dietary carotenoid
concentrations is generally observed after several weeks of adaptation
(Caldern et al., 2007; Nozire et al., 2006b) and d) milk colour index alone
is unable to provide complete discrimination of diet effects (Nozire et al.,
2006a)
Even though no reports have been presented about the relationship
between nutrition and milk appearance, this trait is an important factor
when milk is evaluated by the industry and consumers. No effect on milk
appearance was found among the treatments used in this study (Papers II and
IV).
45
6 Main Findings and Conclusions
Moringa can help small and medium-scale farmers overcome shortages of
good quality feeds and therefore sustain and improve their livestock systems.

Under Tropical Dry Forest conditions and when phosphorus and potassium
are available in the soil, Moringa can maintain high biomass yield over time
but this requires nitrogen to be supplied in sufficient amounts to cover that
removed at harvest.

Moringa ensiled either alone with 10 or 50 g kg
-1
fresh matter molasses
added or in an mixture with Elephant grass can provide acceptable
fermentation patterns and stability after silo opening, while still maintaining
the nutritive value of the silage.

Ensiled Moringa can be fed to dairy cows in large quantities without any
negative effect on nutrient intake or digestibility. Cows fed large quantities
of Moringa silage can produce the same quantity and quality of milk as cows
fed conventional Elephant grass diets.

Moringa leaf meal is a potential source of protein to supplement poor
quality forage such as Elephant grass. It can successfully replace commercial
concentrate constituents for dairy cows as long as the substitution is
isocaloric and isoproteinic.

While a fresh Moringa diet can lead to off-flavour and aroma in milk, a
Moringa silage diet gives milk with good organoleptic characteristics.




46






47
7 Future research
This study gives a good idea of how nitrogen fertilisation can affect
biomass yield and chemical composition over a two-year period, but
experiments over longer periods such as five years can be valuable to
further verify whether high production levels can be maintained at
the planting densities and fertilisation levels recommended here.
Urea was used as a source of nitrogen in this study. However, different
nutrients such as phosphorus and potassium as well different fertiliser
sources, either mineral or organic, should be studied in combination
with different irrigation regimes.
Silage made from combinations of Moringa and Elephant grass or
sugar cane are well documented here. However, combinations with
other common tropical grasses and the use of silage as a supplement
to grazing animal still need to be investigated.
More studies with MLM in other common tropical diets are needed to
obtain detailed results with regard to how feeding Moringa affects
dairy production under conditions where protein quantity and
quality are limiting factors
Further studies are needed on the potential of MLM as an alternative
protein source for milk production based not only on biological
results but also including economic aspects.
It would be valuable to study protein availability of Moringa diets in
vivo.



48


49
References
Afuang, W., Siddhuraju, P., Becker, K. (2003). Comparative nutritional evaluation of raw,
methanol extracted residues and methanol extracts of Moringa (Moringa oleifera Lam)
leaves on growth performance and feed utilization in Nile tilapia (Oreochromis niloticus
L). Aquaculture Research 34, 1147-1159.
Agrodesierto. (2010). Moringa-Moringa oleifera, Programas agroforestales, Retrieved October 11,
2010 from http://www.agrodesierto.com/moringa.html.
Anhwange, B., Ajibola, V., Oniye, S. (2004). Amino acid composition of the seeds of
Moringa oleifera (Lam), Detarium microcarpum (Guill & Sperr) and Bauhinia monandra (Linn.).
Chemistry Classification Journal 9, 13.
AOAC. (1984). Official Methods of Analysis, 14
th
edition. Association of Official Analytical
Chemists. Washington, USA.
AOAC. (2000). Official Methods of Analysis, 17
th
edition. Association of Official Analytical
Chemists. Washington, USA.
APHA. (2001). Compendium of Methods for the Microbiological Examination of Foods. American
Public Health Association, Washington, USA.
Aregheore, E. (2002). Intake and digestibility of Moringa oleifera-batiki grass mixture by
growing goats. Small Ruminant Research 46, 23-28.
Ashbell, G., Weinberg, A., Azrieli, A., Hen, Y., Orbe, B. (1990). A simple system to study
the aerobic determination of silages. Canadian Agricultural Engineering 33, 391-393.
Atega, T., Robles, A., Alinea, C., Lanting, E. (2003). Nutritive value of leaf meals from
fodder trees and shrubs as affected by drying methods. PCARRD Highlights. Philippines.
Atkinson, A., Donev, A. (1992). Optimum experimental designs. Oxford University Press.
Oxford, United Kingdom.
BCN (Banco Central de Nicaragua). (2009). 50 aos de estadsticas macroeconomicas 1960-2009.
Managua, Nicaragua. 139 pp.
Becker, K. (1995). Studies on utilization of Moringa oleifera leaves as animal feed. Intitute of
Animal Production in Tropics and Subtropics 480, 15.
Beer, J. (1989). Experiencias con cercas de rboles forrajeros en Costa Rica y Nicaragua. In:
Beer, J., Fassbender, H., Heuveldop, J. (eds) 1989. Avances en la investigacin agroforestal.
Turrialba, Costa Rica. CATIE-GTZ. 244-253.
50
Belli, R., Tekelenburg, T., Siria, I. (2009). La transicin hacia la sostenibilidad. Reduccin del
impacto futuro del sector ganadero sobre la pobreza y la biodiversidad. EDISA. Managua,
Nicaragua. 150 pp.
Benavidez, J. (1996). Research on Forage Trees. In: Tropical Feeds and Feeding Systems. FAO,
Rome, Italy. 277 pp.
Bendall, J. (2001). Aroma compound of fresh milk from New Zealand cows fed different
diets. Journal of Agriculture and Food Chemistry 49, 4825-4832.
Bennett, R., Mellon, F., Foidl, N., Pratt, J., Dupont, M., Perkins, L., Kroon, P. (2003).
Profiling glucosinolates and phenolics in vegetative and reproductive tissues of the multi-
purpose trees Moringa oleifera L. (Horseradish tree) and Moringa stenopetala L. Journal of
Agricultural and Food Chemistry 51, 3546-3553.
Ben Salem, H., Makkar, H. (2009). Defatted Moringa oleifera seed meal as a feed additive for
sheep. Animal Feed Science and Technology 150, 27-33.
Breme, K., Fernandez, X., Meirhenrich, U., Brevard, H., Joulain, D. (2007). Identification
of new, odor-active thiocarbamates in cress extract and structure Activity studies on
synthesized homologues. Journal of Agricultural and Food Chemistry 55, 1932-1938.
Broderick, G. (2003). Effects of varying dietary protein and energy levels on the production
of lactating dairy cows. Journal of Dairy Science 86, 1370-1381.
Brudzewski, K., Osowski, S., Markiewicz, T. (2004). Classification of milk by means of an
electronic nose and SVM neural network. Sensors and Actuators B: Chemical 98, 291-298.
Buxton, D., Muck, R., Harrison, J. (2003). Silage Science and Technology. American Society of
Agronomy, Madison, Wisconsin.
Caldern, F., Chauveau-Duriot, B., Pradel, P., Martin, B., Graulet, B., Nozire, P. (2007).
Variations in carotenoids, vitamins A and E, and color in cows plasma and milk following
as shift from hay diet to diets containing increasing levels of carotenoids and vitamin E.
Journal of Dairy Science 90, 5651-5664.
Crdenas, J., Sandoval, C., Solorio, F. (2003). Chemical composition of grass and forage trees
mixed silages. Tcnica Pecuaria en Mxico 41, 283-294.
Carvalho, M., Freitas, V., Xavier, D. (2001). Contribuo dos sistemas silvipastoris para a
sustentabilidade da atividade leiteira. EMBRAPA, Gado de Leite, 85-107.
Castro, M., Mosquera-Losada, M., Rigueiro-Rodrguez, A. (1998). Pasture and tree
evolution in silvopastoral systems of Colombia. Forest Ecology and Management 230, 135-
145.
Croissant, A., Washburn, S., Dean, L., Drake, M. (2007). Chemical properties and consumer
perception of fluid milk from conventional and pasture-based production systems. Journal
of Dairy Science 11, 4942-4953.
Dash, S., Gupta, N. (2009). Effect of inorganic, organic and biofertilizer on growth of hybrid
Moringa oleifera (PKM1). Science 4, 630-635.
de Leeuw, P., Omure, A., Staal, S., Thorpe, W. (1998). Dairy production systems in the
tropics: A review. In: Falver, L., Chantalakhan, C. (eds). 1998. Smallholders dairying in the
tropics. Nairobi, ILRI. pp 19-44.
Dongmeza, E., Siddhuraju, P., Francis, G., Becker, K. (2006). Effects of dehydrated
methanol extracts of Moringa (Moringa oleifera Lam) leaves and three of its fractions on
51
growth performance and feed nutrient assimilation in Nile tilapia (Oreochromis niloticus L).
Aquaculture 261, 407-422.
Dougherty, R., Shipe, W., Gudnason, V., Ledford, R., Peterson, R., Scarpellino, R. (1962).
Physiological mechanisms involved in transmitting flavors and odors to milk. I.
Contribution of eructated gases to milk flavor. Journal of Dairy Science 4, 472-476.
Durr, P. (1992). Manual de rboles forrajeros de Nicaragua. Ministerio de Agricultura y
Ganadera. Nicaragua. 125 pp.
El hassan, S., Lahlou, A., Newbold, C., Wallace, R. (2000). Chemical composition and
degradation characteristics of foliage of some African multipurpose trees. Animal Feed
Science and Technology 86, 27-37.
Ella, A., Jacobsen, C., Str, W., Blair, G. (1989). Effect of plant density and cutting
frequency on the productivity of four tree legume. Tropical Grassland 23, 28-34.
Faizi, S., Siddiqui, B., Saleem, R., Siddiqui, S., Aftab, K., Gilani, A. (1995). Fully acetylated
carbamate and hypotensive thiocarbamate glycosides from Moringa oleifera. Phytochemistry
38, 957-963.
Fales, S., Gustine, D., Bosworth, S., Hoover, R. (1987). Concentration of glucosinolates and
S-methylcysteine sulfoxide in ensiled rape (Brassica napus L.). Journal of Dairy Science 70,
2402-2405.
FAO. (1988). FAO/UNESCO soil map of the world, revised legend, with corrections and
updates. World soil resources report 60. FAO, Rome.
FAO. (2006) Production yearbook 2006. Food and Agriculture Organization of the United
Nations, vol 55. Statistica Series, Rome, Italy, 276 pp.
Figueiredo, R., Rodrigues, A., Cu Costa, M. (2007). Volatile composition of red clover
(Trifolium pratense L) forages in Portugal: The influence of ripening stage and ensilage.
Food Chemistry104, 1445-1453.
Foidl, N., Makkar, H., Becker, K. (2001). The potential of Moringa oleifera for agricultural
and industrial uses. What development potential for Moringa products? Dar Es Salaam.
Fujihara, T., Kakengi, A., Shem, M., Sarwatt, S. (2005). Can Moringa oleifera be used as a
protein supplement for ruminants? Asian Australasian Journal of Animal Science 18, 42.
Fujisaka, S., Holmann, F., Peters, M., Schimdt, A., White, D., Burgos, C., Ordoez, J.,
Mena, M., Posas, M., Cruz, H., Davis, C., Hincapi, B. (2005). Estrategias para
minimizar la escasez de forrajes en zonas con sequas prolongadas. Honduras y Nicaragua.
Documento de trabajo No 21. Centro Internacional de Agricultura Tropical. Cali. Colombia. 28
pp.
Gidamis, A., Panga, J., Sarwatt, S., Chove, B., Shayo, N. (2003). Nutrients and antinutrients
contents in raw and cooked leaves and mature pods of Moringa oleifera Lam. Ecology and
Food Nutrition 42, 1-123.
Ghosh, M., Atreja, P., Buragohain, R., Banyopadhyay, S. (2007) Influence of short-term
Leucaena leucocephala feeding on milk yield and its composition, thyroid hormones,
enzyme activity, and secretion of mimosine and its metabolites in milk cattle. Journal of
Agricultural Science 145, 407-414.
Hammond, A. (1995). Leucaena toxicosis and its control in ruminants. Journal of Animal
Science 73, 1478-1492.
52
Hartley, S., Nelson, K., Gorman, M. (1995). The effect of fertilizer and shading on plant
chemical composition and palatability to Okney voles, Microtus arvalis orcadensis. Oikos 72,
79-87.
Holden, L. (1999). Comparison of methods of in vitro dry matter digestibility for ten feeds.
Journal of Dairy Science 82, 1791-1794.
Holmann, F., Argel, P., Rivas, L., White, D., Estrada, R., Burgos, C., Prez, E., Ramrez,
G., Medina, A. (2004). Vale la pena recuperar pasturas degradadas? Una evaluacin de los
beneficios y costos desde la perspective de los productores y extensionistas pecuarios en
Honduras. International Livestock Research Institute. ILRI.
Kaimowitz, D. (1995). Livestock and deforestation in Central America. Documento de
discusin no 9. IICA. Costa Rica.
Kakengi, A., Kaijage, J., Sarwatt, S., Mutayoba, S., Shem, M., Fujihara, T. (2007). Effect of
Moringa oleifera leaf meal as a substitute for sunflower seed meal on performance of laying
hens in Tanzania. Livestock Research for Rural Development 19, 120.
Kalac, P. (2010). The effect of silage feeding on some sensory and health attributes of cows
milk: a review. Food Chemistry. DOI 10.1016/j.foodchem.2010.08.077
Lindgren, E. (1979). Vallfodrets nringsvrde bestmt in vivo och med olika
laboratoriemetoder. Report 45. The Department of Animal Nutrition and Management.
The Swedish University of Agricultural Sciences, Uppsala, Sweden.
Lindgren, S. (1990). Microbial dynamics during silage fermentation. In: Lindgren, S.,
Lundn, K. (1990). Proceedings of the Eurobac conference 1986. Grass and Forage Report.
SLU. Uppsala, Sweden.
MacGregor, C. (2000). Directory of Feeds and Feed Ingredients. Hoards Dairyman Milwaukee
USA. pp 95.
Madalla, N. (2008). Novel feed ingredients for Nile tilapia (Oreochromis niloticus L). Doctoral
thesis. University of Stirling. Scotland, UK. 212 pp.
MAGFOR (Ministerio de Agricultura, Ganadera y Forestal). 2009. Informe anual de produccin
agropecuaria ciclo agrcola 2008-2009 y periodo pecuario 2009. Managua, Nicaragua. 89 pp.
Makkar, H., Becker, K. (1997). Nutrients and antiquality factors in different morphological
parts of the Moringa oleifera tree. Journal of Agricultural Science 128, 311-332.
Manh, L., Nguyen, N., Ngoi, T. (2005). Introduction and evaluation of Moringa oleifera for
biomass production and feed for goats in the Mekong delta. Livestock Research for Rural
Development 17, 9.
Martin, B., Verdier-Metz, I., Buchin, S., Hurtaud, C., Coulon, J. (2005). How do the nature
of forages and pasture diversity influence the sensory quality of dairy livestock products?
Animal Science 81, 205-212.
Martin, B., Hurtaud, C., Graulet, B., Ferlay, A., Chilliard, Y., Coulon, J. (2009). Herbe et
qualities nutritionnelles et organoleptiques des produits laitiers. Fourrages 199, 291-310.
Mauricio, R., Sousa, L., Moreira, G., Reis, G., Gonalves, L. (2008). Silvopastoral systems as
a sustainable alternative to animal production in the tropics. In: Casteln, O., Bernus, A.,
Ruiz, R., Mould, F. (Eds) 2008. Opportunities and Challenges for Smallholder Ruminant
Systems in Latin America. Resources management, food safety, quality and market access.
Toluca, Mexico. 187-200.
McDonald, P., Edwards, R., Greenhalgh, J. (1988). Animal Nutrition. Longman Group, UK.
53
McDonald, P., Henderson, N., Heron, S.(2002). Biochemistry of Silage. Chalcombe
Publications, Marlow, UK
Mendieta-Araica, B., Reyes-Snchez, N., Alfranca, O. (2000). Estrategia de desarrollo
pecuario para el departamento de Chontales, Nicaragua. Thesis Magister Science. Universitat
Autonoma de Barcelona. Spain.
Mendieta-Araica, B., Sprndly, E., Reyes-Snchez, N., Norell, L., Sprndly, R. (2009).
Silage quality when Moringa oleifera is ensiled in mixtures with Elephant grass, sugar cane
and molasses. Grass and Forage Science 64, 364-373.
Mommer, L., Lensen, J., Huber, H., Visser, E., de Kroon, H. (2006). Ecophysiological
determinants of plant performance under flooding: a comparative study among seven
plant families. Journal of Ecology 94, 1117-1129.
Mosquera-Losada, M., Riguiero-Rodrguez, A., McAdam, J. (2005). Silvopastoralism and
sustainable land management.
Murro, J., Muhikambele, V., Sarwatt, S. (2003). Moringa oleifera leaf meal can replace
cottonseed cake in the concentrate mix fed with Rhodes grass (Chloris gayana) hay for
growing sheep. Livestock Research for Rural Development 15,11.
Mustapha, Y., Babura, S. (2009). Determination of carbohydrate and -carotene content of
some vegetables consumed in Kano metropolis, Nigeria. Bayero Journal of Pure and Applied
Sciences 2, 119-121.
Nash, J. (1985). Crop Conservation and Storage in Cool, Temperate Climates. Pergamon Press,
Oxford UK.
Nouala, F. (2004). Comparison of plant cell degrading community in the rumen of NDama
and NDama*Jersey crossbred cattle in relation to in vivo and in vitro cell wall degradation.
PhD Thesis. University of Hohenheim.
Nouala, F., Akinbamijo, O., Adewumi, A., Hoffman, E., Muetzel, S., Becker, K. (2006).
The influence of Moringa oleifera leaves as substitute to conventional concentrate on the in
vitro gas production and digestibility of groundnut hay. Livestock Research for Rural
Development 18, 9.
Nozire, P., Grolier, P., Durand, D., Ferlay, A., Pradel, P., Martin, B. (2006a). Variations in
carotenoids, fat-soluble micronutrients, and color in cows plasma and milk following
change in forage and feeding level. Journal of Dairy Science 89, 2634-2648.
Nozire, P., Graulet, B., Licas, A., Martin, B., Grolier, P., Doreau, M. (2006b). Carotenoids
for ruminants: from forages to dairy products. Animal Feed Science and Technology 131,
418-450.
NRC. (1988). Nutrient Requirements of Dairy Cows. National Academy of Science.
Washington, USA.
Nuhu, F. (2010). Effect of Moringa leaf meal (Molm) on nutrient digestibility, growth,
carcass and blood indices of weaner rabbits. MSc Thesis. Kwame Nkrumah University of
Science and Technology. Ghana.
Nursten, H. (1997). The flavor of milk and dairy products: I. Milk of different kinds, milk
powder, butter and cream. International Journal of Dairy Technology 50, 48-56.
Oerlemans, K., Bosh, C., Vernerk, R., Dekker, M. (2006). Thermal degradation of
glucosinolates in red cabbage. Food Chemistry 95, 19-29.
54
Oldham, J. (1984). Protein-energy interrelationships in dairy cows. Journal of Dairy Science 67,
1090-1114.
Oliveira, J., Souto, J., Santos, R., Souto, P., Maior Jnior, S. (2009). Adubao com
diferentes estercos no cultivo de Moringa (Moringa oleifera LAM.). Revista Verde de
Agroecologia e Deselvolvimiento Sustentvel. 4.
Olugbemi, T., Murayoba, S., Lekule, F. (2010). Effect of Moringa (Moringa oleifera) inclusion
in Cassava based diets fed to broilers chickens. International Journal of Poultry Science 9, 363-
367.
Olsson, L., Wilgert, E. (2007). Moringa oleifera- an evaluation of drying methods to produce
dry season feed for cattle. Minor Field Studies No 390. External Relations. Swedish
University of Agricultural Sciences. Uppsala, Sweden.
Pamo, E., Boukila, B., Tonfack, L., Momo, M., Kana, J., Tendokeng, F. (2005). Influence
de la fumure organique, du NPK et du mlange des deux fertilisants sur la croissance de
Moringa oleifera Lam. Dans lOuest Cameroun. Livestock Research for Rural Development. 17.
Panciera, M., Kunkle, W., Fransen, S. (2003). Minor silage crops. In: Buxton, D., Muck, R.,
Harrison, J. (eds) 2003. Silage science and technology. American Society of Agronomy, Inc; Crop
Science Society of America, Inc.; Soil Science society of America, Inc. Madison, USA. 781-824.
Patterson, H., Lucas, H.(1962). Change-over design. Technical Bulletin 147. North Carolina.
USA.
Pedroso, A., Nussio, L., Loures, D., Paziani, S., Santana, D., Igarasi, M., Coelho, R., Packer,
H., Horii, J., Gomes, L. (2005). Fermentation and epiphytic microflora dynamics in sugar
cane silage. Scientia Agricola 62, 427-432.
Pereira, F. (1988). Alimentos, Manual de anlisis fsico qumico. Universidad Autnoma de
Yucatn. Direccin General de Difusin y Comunicacin. Mrida, Yucatn, Mxico. 190 pp.
Phengvichit, V., Ledin., S., Horne, P., Ledin, I. (2006). Effects of different fertilizers and
harvest frequencies on foliage and tuber yield and chemical composition of foliage from
two cassava (Manihot esculenta, Crantz) varieties. Tropical and Subtropical Agroecosystems 6,
177-187.
Pok, S., Bun, Y., Ly, J. (2005). Physico-Chemical properties of tropical tree leaves may
influence its nutritive value for monogastric animal species. Revista computarizada de
produccin porcina 12, 31-34.
Reyes, O., Fermin, A. (2003). Terrestrial leaf meals or freshwater aquatic fern as potential
feed ingredients for farmed abalone Haliotis asinine (Linnaeus 1758). Aquaculture Research
34, 593-599.
Reyes-Snchez, N., Ledin, S., Ledin, I. (2006a). Biomass production and chemical
composition of Moringa oleifera under different management regimes in Nicaragua.
Agroforestry Systems 66, 231-242.
Reyes-Snchez, N., Sprndly, E., Ledin, I. (2006b). Effect of feeding different levels of
foliage of Moringa oleifera to creole dairy cows on intake, digestibility, milk production and
composition. Livestock Science 1001, 24-31.
Reyes-Snchez, N., Mendieta-Araica, B., Faria, T., Mena, M. (2008). Gua de
suplementacin estratgica para bovinos en poca seca. Serie Guas Tcnicas 12,
Universidad Nacional Agraria, Managua, Nicaragua.
55
Richter, N., Siddhuraju, P., Becker, K. (2003). Evaluation of nutritional quality of Moringa
(Moringa oleifera Lam.) leaves as an alternative protein source for Nile tilapia. Aquaculture
217, 599-611.
Rodrguez, C., Arias, R., Quiones, J. (1994). Efecto de la frecuencia de poda y el nivel de
fertilizacin nitrogenada, sobre el rendimiento y calidad de la biomasa de Morera (Morus
spp.) en el trpico seco de Guatemala. In: Benavidez, J (ed). (1994). Arboles y arbustos
forrajeros en Amrica central. Vol 2. Serie Tcnica # 236. CATIE pp 515-529.
Salmern-Miranda, F., Bath, B., Eckerten, H., Forkman, J., Wisvtad, M. (2007).
Aboveground nitrogen in relation to estimated total plant uptake in maize and bean.
Nutrient Cycling Agroecosystems 79, 125-139.
Snchez, P., Speed, S.,(1999). Assessing the adoption potential of silvopastoral practices in
Latin America. Agricultural Systems 38, 17-43.
Snchez-Machado, D., Nez-Gastlum, J., Reyes-Moreno, C., Ramrez-Wong, B., Lpez-
Cervantes, J. (2010). Nutritional quality of edible parts of Moringa oleifera. Food Analytical
Methods 3, 175-180.
Sarwatt, S., Milangha, M., Lekule, F., Madala, N. (2004). Moringa oleifera and cottonseed
cake as supplements for smallholders dairy cows fed Napier grass. Livestock Research for
Rural Development 16, 6.
SAS. (2004). User guide, version 9.1.2. Statistical Analysis System Institute Inc. USA.
Schingoethe, D. (1996). Dietary influence on protein level in milk yield in dairy cows.
Animal Feed Science and Technology 60, 181-190.
Shipe, W., Ledford, A., Peterson, R., Scanlan, R., Geerken, H., Dougherty, R., Morgan,
M. (1962). Physiological mechanisms involved in transmitting flavors and odors to milk.
II. Transmission of some flavor components of silage. Journal of Dairy Science 4, 477-480.
Soliva, C., Kreuzer, M., Foidl, N., Foidl, G., Machmller, A., Hess, H. (2005). Feeding
value of whole and extracted Moringa oleifera leaves for ruminants and their effects on
ruminal fermentation in vitro. Animal Feed Science and Technology 118, 47-62.
Solorio-Snchez, F., Armendariz-Yaez, I., Ku-Vera, J. (2000). Chemical composition and
in vitro dry matter digestibility of some fodder trees from South-east Mexico. Livestock
Research for Rural Development 12, 4.
Steinfeld, H., Gerber, P., Wassenaar, T., Castel, V., Rosales, M., Haan, C. (2006). Livestocks
long shadow. Environmental Issues and Options. Rome, FAO.
Sultan, J., Ur-rahim, I., Nawaz, H., Yaqoob, M., Javed, I. (2008). Nutritional evaluation of
fodder tree leaves of northern grassland of Pakistan. Pakistan Journal of Botany 40, 2503-
2512.
Sutton, J. (1989). Altering milk composition by feeding. Journal of Dairy Science 72, 2801-
2814.
Thomas, E. (1981). Trends in milk flavors. Journal of Dairy Science 64, 1023-1027.
Tjandraatmadja, M., Norton, B., Macrae, I. (1994). Ensilage of tropical grasses mixed with
legumes and molasses. World Journal of Microbiology and Biotechnology 10, 82-87.
USDA (United States Department of Agriculture), Natural Resources Conservation Service.
(1995). Soil survey laboratory information manual. Natural Soil Survey Center, Soil Survey
Laboratory. SSIR 45.
56
Van Soest, P., Robertson, J., Lewis, B. (1991). Methods for dietary fibre, neutral-detergent
fibre and non-starch polysaccharides in relation to animal nutrition. Journal of Dairy Science
74, 3583-3597.
Vipond, J., Duncan, A., Turner, D., Goddyn, L., Horgans, G. (1998). Effects of feeding
ensiled kale (Brassica oleracea) on the performance of finishing lambs. Grass and Forage
Science 53, 346-352
Walker, N., Gray, I. (1970). The glucosinolate of Land Cress (Coronopus didymus) and its
enzymic degradation products as precursors of off-flavor in milk A review. Journal of
Agriculture and Food Chemistry 18, 346-352.
Weissbach, F. (1996). New developments in crop conservation. XI International Silage
Conference. IGER, Aberystwyth, UK, pp 11-25.
Witting de Penna, E. (1981). Evaluacin sensorial, una metdica que mide calidad. II.
Evaluacin mediante el test de valoracin con escala de Karlsruhe. Alimentos 6, 25-31.
WHO&FAO (World Health Organization and Food and Agriculture Organization). (2007).
Milk and milk products. Codex alimentarius. Rome. pp 242.




57
Acknowledgements
This thesis is the result of great team work. There were so many persons
involved that is impossible for me to remember all of them; however, they
all will always have my gratitude.
I would like first to thank the Swedish International Development
Agency SIDA for the financial support not only for this work but for the
more than three decades of helping my country in so many different ways.
Special thanks should be given to Lars Ohlander, who was the pillar
supporting our programme throughout the years.
In Sweden there are so many people who gave me selfless help, support,
advice and friendship. Eva Sprndly at the head of them, my main
supervisor and the person who guided me step by step in this amazing world
of science. She also opened the doors of her home to me and together with
her husband Rolf Sprndly, who is also my supervisor, gave me long and
absolutely interesting conversations as well very delicious meals in our
soirees in this home far away from home; they have been with me beyond
any limit and for them my eternal gratitude. Magnus Halling, Brje Ericson,
Tommy Pauly, Margarita Cuadra and her husband Peter Petersson, Gloria
Gallardo, Cristian Alarcon, Edwin Briceo and his wife Petra and Jette
Jacobsen were always there when I needed them, and this long and lonely
walk was easier because of them.
To the staff of the Animal Nutrition and Management Department my
endless gratitude for their permanent willingness to share knowledge, to
Brian Ogle, Inger Ledin, Margareta Norinder, Jan Erik Lindberg, Peter
Udn, Ana-Greta Haglund, Torbjrn Lundh, Karin Lyberg, Ragnar
Tauson, Maria Neil, Hans Petterson, Ewa Wredle, Sigrid Agens and
Gunnar Petterson.
Who knows better than my fellow PhD students all the endurance
needed in this adventure, but together we managed to end up with success,
58
to Malavanh, Seuth, Vo Lam, Millogo, Martin, Markus, Emma, Hue,
Haoyu, Lampheuy, Daovy, Thi Da, Maja, Ulises, Cicci, Mikaela, Roldan,
Odum, Gerardo, Matilde, Laki, Vanthong, Teddy, Thieu, Salimata, Thuy,
Tram, Salifou and so many more, thanks a lot for your sincere friendship.
All my life, part of you will always stay with me.
In Nicaragua I would like to express my sincere gratitude to Nadir
Reyes, my national supervisor, who has also been my friend for more than
15 years, his continued support and encouragement finally convinced me to
take the decision to apply for PhD studies. My gratitude also to my brother
Allan Bez, we share no blood but spirit and throughout more than 30 years
he has been giving me his friendship. Friends, thank you.
This study would not have been possible without the tireless work and
full dedication of my research team: Rosario Rodrguez, Zoila Romero,
David Aruz, Vctor Acevedo, Elio Zeledn, Marlon Aragn, Nstor
Hurtado and his brother, Jasser Garca, Vera Argello, Ivan Oliva, Perla
Gutirrez and Eliezer Pichardo. No matter whether the work was done
under the strong tropical sun or under the heavy rain harvesting Moringa or
during our endless, sleepless nights collecting faeces or measuring
temperature in silages, they always kept their optimism and commitment
generating useful knowledge to help our farmers. To all of you, thank you
very much.
In Nicaragua there are also other persons I want to thank because they
did everything in their power to support me; Pablo Valdivia put his cows
and his farm in my hands to perform my research and helped me as much as
possible; Miguel Ros, Norlan Caldera, Vctor Aguilar, Edgardo Jimnez,
Juan Lpez and his staff, Leonardo Garca and Vidal Marn helped me either
with animals, forage, analysis or advice any time I came to them. Lester
Rocha and Francisco Salmern continuously gave me advice and support
with my experiments and papers.
Last but never least, I deeply want to thank my family, this achievement
is also their achievement. I stole from them the time I dedicated to my
studies but still they always give me their love, support and encouragement.
If I have reached the end it is because they also sacrificed many things for
me and for that there are no words in the entire universe to express how
much I want to thank them. Ivania, Briannita, Bryan see dedication for you,
all my love.

1

Biomass production and chemical composition of Moringa
oleifera under different planting densities and levels of nitrogen
fertilization
B NenuietaAiaica
a
E Spoinuly
a
N ReyesSnchez
c
F SalmeionNiianua
u
N Balling
b

a
Bepaitment of Animal Nutiition anu Nanagement Sweuish 0niveisity of Agiicultuial Sciences P0 Box
0ppsala Sweuen
b
Bepaitment of Ciop Piouuction Ecology Sweuish 0niveisity of Agiicultuial Sciences P0 Box
0ppsala Sweuen
c
Facultau ue Ciencia Animal 0niveisiuau Nacional Agiaiia P0 Box Nanagua Nicaiagua
u
Facultau ue Agionomia 0niveisiuau Nacional Agiaiia P0 Box Nanagua Nicaiagua
Coiiesponuing authoi P0 Box tel Fax Email auuiess
magnushallingsluse
Abstract
The effect of uiffeient planting uensities B anu B plants ha

anu levels of
nitiogen feitilization N N N anu N kg N ha

yeai

on biomass piouuction anu

chemical composition of Moringa oleifera was stuuieu in a splitplot uesign with foui ianuomizeu
complete blocks ovei two yeais with eight cuts yeai

B piouuceu significantly highei total uiy mattei

yielu TBNY anu fine fiaction yielu FFBN anu ton ha

iespectively compaieu with
anu ton ha

foi B uiowth iate in B was highei than in B compaieu with ton ha

uay

Aveiage plant height was cm iiiespective of planting uensity Feitilization at the N anu N
levels piouuceu the highest TBNY anu FFBN in both yeais of the stuuy anu along all cuts The
inteiaction between cut anu yeai was significant with the highest TBNY anu FFBN uuiing the iainy
season in the seconu yeai Chemical composition of fiactions showeu no significant uiffeiences
between planting uensities Significantly highei ciuue piotein content was founu in the coaise fiaction
at feitilizei levels N anu N anu g kg

BN compaieu with the lowei levels The iesults

inuicate that Noiinga can maintain up to ton ha

uiy mattei yielu unuei uiy tiopical foiest

conuitions ovei time at a planting uensity of plants ha

if the soil is iegulaily supplieu with N

2

at a level of appioximately kg ha yeai

in conuitions wheie phosphoius anu potassium aie not

limiting
Keyword: Moringa, fertilization, biomass production, chemical composition, planting density
Introduction
Nany ieseaicheis have shown an incieasing inteiest in a silvopastoial appioach with tiees anu shiubs
as alteinative fouuei to cattle in tiopical iegions The impioveu feeu supply foi animals ovei the yeais
in silvopastoial systems is iepoiteu to allow faimeis to oveicome the ueleteiious effects of seasonality
in theii piouuction anu theii income Aganga anu Tshwenyane Bolmann anu Lascano
Leng Smith Zamoia et al Theie aie many tiees anu shiubs of inteiest in
agiofoiestiy systems anu one inteiesting tiee species that has ieceiveu a gieat ueal of attention uuiing
iecent yeais is Moringa oleifera Lam synonym M pterygosperma uaeitn M moringa Nills This species
commonly iefeiieu to as Noiinga oi uiumstick tiee is a meuiumsizeu tiee that is inuigenous in
noithwest Inuia anu now wiuely uistiibuteu thioughout the tiopics This multipuipose tiee has been
cultivateu foi use as foou meuicine anu feeu Anwai et al 0liveiia et al ReyesSnchez et
al. a It belongs to the monogeneiic family Noiingaceae anu is uistinguisheu by its usually
tiipinnate leaves with leaflets mm long anu petioles yellow oi white without ieu stieaks
Ramachanuian et al. It can be piopagateu eithei by uiiect seeuing oi haiu stem cuttings
Noiton Fuitheimoie it contains a negligible amount of antinutiitional factois has a high
concentiation of ciuue piotein CP anu significant contents of vitamins A B anu C in the foliage
Feiieiia et al Nakkai anu Beckei This tiee can thiive in subtiopical anu tiopical
climates Buke It can giow in a wiue iange of soils Noiton but it giows best on a uiy
sanuy soil Buke Noiinga has been iepoiteu to be uiought toleiant Buke Noiton
but piolongeu uiy peiious iesult in loss of leaves ReyesSnchez et al b
Tests on uiffeient cutting fiequencies of Noiinga have shown that a uay cutting inteival is the most
appiopiiate when biomass is to be useu as animal feeu Calub Foiul et al ReyesSnchez et
al b The effect of planting uensity on yielu of uiy mattei BN has also been stuuieu anu Foiul et
al iepoiteu incieasing BN yielus fiom to ton ha

at anu plants ha

iespectively Bowevei uue to the high moitality at veiy high planting uensities those authois
iecommenueu plants ha

as the optimum In anothei stuuy ReyesSnchez et al b

iepoiteu an aveiage BN yielu of ton ha

uuiing the fiist yeai of an expeiiment with no significant

uiffeiences among thiee uiffeient planting uensities anu plants ha

Although high uensities aie positively coiielateu with high BN yielus the spatial aiiangement in the
3

fielu the high amount of laboui neeueu anu uifficulties uuiing haivesting make high uensities
impiactical foi small anu meuiumscale faimeis Theiefoie stuuies on lowei uensities moie auapteu to
the piactical neeus of small anu meuiumsizeu faims aie still neeueu
In agiofoiestiy systems wheie high amounts of nutiients aie iemoveu fiom the plantation aiea at
haivest feitilization in a geneial sense seems to be the only way to maintain sustainable piouuction
Szott anu Kass This was veiy eviuent in the iesults fiom a twoyeai stuuy piesenteu by Reyes
Snchez et al b wheie BN yielu of unfeitilizeu Noiinga uecieaseu by about uuiing the
seconu yeai of piouuction In spite of this few feitilization expeiiments have been peifoimeu with
Noiinga unuei fielu conuitions Bash anu uupta 0liveiia et al Pamo et al
The aim of this stuuy was theiefoie to ueteimine the effect of two plant uensities anu foui levels of
nitiogen feitilization on biomass piouuction anu chemical composition of leaves petioles anu stems of
Moringa oleifera unuei fielu conuitions
Material and methods
Location of experimental area
The stuuy was peifoimeu uuiing two consecutive yeais at the National Agiaiian 0niveisity 0NA
faim in Nanagua Nicaiagua N W altituue m above sea level This is an
ecological zone of uiy tiopical foiest with mean annual iainfall of mm anu ielative humiuity
INETER The aiea has a uiy season fiom Novembei to Apiil anu a iainy season fiom Nay to
0ctobei Bowevei in the miuule of the iainy season }ulyAugust theie is a shoit peiiou with ieuuceu
iainfall known as Canicula which usually begins in the last week of }uly anu enus in miuAugust
Nean annual tempeiatuie is C with the highest tempeiatuies occuiiing towaiu the enu of Apiil
This stuuy was conuucteu as a fielu expeiiment fiom }une to 0ctobei in an aiea which was
pieviously fallow lanu Yeai one was peifoimeu in the uiy season of anu the iainy season of
while yeai two was the uiy season of anu the iainy season of Theie weie gieat
vaiiations in piecipitation between yeais Figuie with iainfall of anu mm in anu
but with unusually low iainfall uuiing Nay anu }une of anu veiy uiy conuitions in
mm
4


Fig. 1 Nonthly piecipitation mm anu tempeiatuie C in Nanagua Nicaiagua uuiing the peiiou }une
to Becembei
The soil of the expeiimental aieas was classifieu as slightly alkaline with pB The peicentage of
oiganic mattei anu nitiogen N in the soil was anu iespectively with ppm of
available phosphoius P anu meq g

soil of available potassium K The soil belongs the

taxonomic oiuei of Anuosols FA0 anu clay loam textuial class 0SBA It has goou
uiainage anu can be consiueieu suitable foi agiicultuie
Soil preparation and sowing
Soil piepaiation was uone by conventional tillage using a tiactoi anu mechanical tools to cleai the lanu
of plant uebiis followeu by uisc ploughing uisc cultivation anu two haiiowings 0ntieateu seeus of
Noiinga weie useu foi piopagation In }une seeus weie sown in cm ueep holes at the stuuy site
seeus pei hole Aftei months of giowth the stanu was thinneu anu only one healthy plant was
kept Iiiigation was not applieu Weeus weie contiolleu manually uays aftei geimination anu eveiy
seconu month thioughout the expeiiment
5

Attack by teimites Heterotermes aureus Snyuei was obseiveu in miu anu ENuE0

SC was
applieu accoiuing the manufactuieis iecommenuations
Fertilization
In this longteim expeiiment }une anu 0ctobei weie set as the feitilization occasions Figuie shows
the uoses of N feitilizei useu in those months anu at sowing To establish suitable expeiimental levels
of Nfeitilization uata fiom a piouuction expeiiment peifoimeu in Nicaiagua with Moringa oleifera
ReyesSnchez et al b weie useu to estimate expecteu BN yielus anu chemical composition of
Noiinga foliage The total ciuue piotein CP yielu on a BN basis iepoiteu by ReyesSnchez et al.
b was uiviueu by to get the amount of N iequiieu by the ciop anu in the piesent
expeiiment this was set as N Levels below anu above this level weie set as N anu N A contiol
tieatment with no feitilization was set as N The levels of N feitilization useu in this stuuy weie
theiefoie N N N anu N kg N ha

yeai

In cuts in 0ctobei anu }une of each

expeiimental yeai N feitilizei was applieu weeks aftei piuning anu was uiiectly incoipoiateu into
the soil by manual hoeing

Fig 2. Boses of N feitilizei applieu to the expeiimental fielu on uiffeient feitilization occasions ovei time
The aim of the feitilization stiategy was to allow the effect of uiffeient N levels to be stuuieu without
iisking P anu K ueficiency that coulu influence the iesults Theiefoie foi calculation of P anu K
iequiiements the uata fiom ReyesSnchez et al b was also useu as iefeience foi estimating uiy
mattei BN yielus anu chemical composition of Noiinga foliage Baseu on the P anu K contents in BN
6

the total amount of those mineials was calculateu anu set as of theii iequiiements but of
the total K anu P iequiieu was applieu at sowing in the entiie expeiimental aiea to ensuie the
iequiiements woulu be met which coiiesponueu to kg of P ha

anu kg K ha

iespectively
0iea tiiple supeiphosphate anu muiiate of potash weie useu as the souice
of N P anu K iespectively
Spatial arrangement
The expeiimental aiea was set up in a fielu using half the aiea m

foi planting anu the iemaining

m

foi a boiuei aiea m wiue stiips between blocks anu m between subplots to facilitate
management of the expeiiment The inuiviuual subplot size was m

anu the net aiea useu foi

haivest was m

to eliminate euge effects 0nifoim spacing between iows m was useu in all
plots With the aim of obtaining two uiffeient planting uensities anu plant ha

anu m spacings between plants weie useu within iows
Experimental design and sampling procedures
The expeiiment hau a splitplot uesign with foui ianuomizeu complete blocks The two planting
uensities weie the main plot factoi anu the foui levels of N feitilization weie the subplot factoi In
total plots weie establisheu
At the stait of the stuuy in 0ctobei the whole plantation was unifoimly cut to a height of cm
above the giounu anu all foliage was iemoveu but not weigheu The iegiowth was haivesteu
thioughout the two subsequent yeais staiting fiom miuNovembei Eveiy uays the iegiowth
was haivesteu at cm height above the giounu using a machete The fiesh biomass fiom each plot
was then weigheu anu iecoiueu to estimate fiesh mattei yielu The mateiial obtaineu fiom each plot
was sepaiateu into two fiactions a fine fiaction which incluueu leaves petioles anu soft stems mm
in uiametei oi smallei anu a coaise fiaction which incluueu stems laigei than mm in uiametei The
weight of each fiaction was iecoiueu anu samples of the fine anu coaise fiaction weie taken foi
subsequent chemical analysis 0bseiveu vaiiables weie total uiy mattei yielu TBNY anu fine
fiaction yielu in BN FFBN both measuieu in metiic tonnes ton ha

yeai

plant height cm anu

giowth iate ton BN ha

uay

Eight cuts pei yeai weie peifoimeu so staiting with the fiist iegiowth
haivest haivesting occasions weie numbeieu fiom one to eight foi each yeai The teim Cut is
hencefoith useu foi these haivesting occasion with Cuts taking place in the uiy season Novembei
Apiil anu Cuts in the iainy season Nay0ctobei of each expeiimental yeai
7

Aveiage height of the plants was estimateu by measuiing the height of five ianuomly selecteu plants in
each subplot of each tieatment Toleuo anu ShulzeKiaft The measuiements weie maue
between the plant base giounu anu the highest tip of the leaves Foi giowth iate the uaily biomass
piouuction ton BN ha

uay

uuiing each giowth peiiou was estimateu using the equation

uiowth iate BN yielu ton ha

Cut

cutting inteival uays

Noitality was calculateu as peicentage of plants that uieu at the enu of the yeai uiviueu by the numbei
of live plants at the beginning of each expeiimental yeai foi each subplot
Chemical analysis
All samples weie uiieu in a foiceu uiaft oven at C foi h Biieu samples weie milleu to pass
thiough a mm sieve foi chemical evaluation Ash anu BN weie analyzeu using A0AC
pioceuuies Each yeai a uetaileu chemical analysis was peifoimeu on the samples collecteu on the
occasion that gave the highest uiy mattei yielu uuiing the yeai Cut The ciuue piotein CP
concentiation was ueteimineu using the Kjeluahl methou A0AC Neutial ueteigent fibie NBF
anu aciu ueteigent fibie ABF contents weie analyzeu as uesciibeu by van Soest et al using
souium sulphite
Statistical analysis
A iepeateu measuies analysis of vaiiance was conuucteu to ueteimine the effect of plant uensity anu
level of feitilization on the vaiiables measuieu using the Nixeu pioceuuie in SAS SAS Tuikeys
paiiwise compaiison pioceuuie was useu The statistical mouel was
Yijklm ijkik lilkl ambjmeijklm with i j k l m
wheie is the oveiall mean i is the fixeu effect of yeai j the fixeu effect of planting uensity k the
fixeu effect of feitilization level ik the inteiaction between yeai anu feitilization level l the fixeu
effect of Cutil the inteiaction between yeai anu Cut kl the inteiaction between feitilization
level anu Cut am the ianuom effect of block bjm the ianuom effect of main plot anu eijklm the ianuom
iesiuual eiioi
0bseivations fiom the same subplot weie assumeu to be coiielateu with a fiist oiuei autoiegiessive
stiuctuie Begiees of fieeuom weie estimateu using the Kenwaiu anu Rogeis methou The inteiactions
planting uensityyeai planting uensityfeitilization level planting uensityCut anu the highei oiuei
inteiactions weie testeu but then iemoveu fiom the mouel uue to lack of significance P
8

Results
Biomass production
The effect of the two planting uensities on TBNY FFBN giowth iate uR anu plant height B aie
shown in Table Significant uiffeiences P weie founu between planting uensities with iegaiu
to TBNY FFBN anu uR An inciease of anu was obseiveu foi TBNY FFBN anu uR
iespectively at plants ha

B compaieu with plants ha

B uuiing the seconu

yeai many plots uiu not piouuce a coaise fiaction No uiffeiences between planting uensities weie
founu in teims of plant height which aveiageu cm
Table Piouuction uata on Moringa oleifera at two uiffeient planting uensities in Nanagua Nicaiagua
Biomass piouuction ton uiy mattei BN ha

giowth iate ton BN ha

uay

anu aveiage plant height

cm least squaie means LSN
vaiiable Planting uensity plants ha

value

Sqit

SE

value

Sqit

SE

Biy mattei yielu
Total
b

a

Fine fiaction
b

a

uiowth iate
b

a

Beight

Squaie ioot Sqit tiansfoimeu uata useu in the statistical analysis of biomass piouuction anu giowth iate to obtain noimal
uistiibution value of tiansfoimeu uata in table is the squaieu value of the LSN obtaineu in the statistical analysis

SE Stanuaiu eiioi
ab
Neans in the same iow with uiffeient supeisciipt uiffei significantly P
In both yeais theie weie no significant uiffeiences in TBYN between N anu N which gave an aveiage
piouuction of anu ton ha

yeai

in yeais one anu two iespectively Bowevei yielus with

both N anu N weie significantly highei P than with N anu N Buiing the TBNY
foi N anu N was anu ton ha

yeai

iespectively while the coiiesponuing values foi the same

tieatments in weie anu ton ha

yeai

N showeu no significant uiffeiences between

yeais but a significant P ieuuction in TBNY of was obseiveu in N uuiing
compaieu with the pievious yeai
9


Y1= 20072008 Y2=20082009
N1=0 N2=278 N3=521 N4=782
Fig 3 Effects of inteiactions between foui uiffeient levels of nitiogen feitilization kg ha

yeai

in two
consecutive yeais on a fine fiaction uiy mattei yielu b giowth iate anu c plant height
10

Significant inteiactions between yeai anu feitilization level iegaiuing FFBN anu uR aie piesenteu
giaphically in Figuie Levels N anu N gave highei values in compaieu with the following
yeai foi FFBN anu uR while N anu N gave an inciease in compaieu with the pievious
yeai foi these vaiiables No inteiactions weie founu between yeai anu feitilization level foi plant
height Plants weie consistently highei in than in the following yeai When the ielationship
between cut anu biomass yielu was analyzeu a maximum of about ton ha

TBNY was obseiveu in

Cut which took place in Nay in both yeais FFBN weie veiy similai foi the fiist five cuts but
significant uiffeiences P weie founu in the inteiactions between yeais anu cuts foi Cut with
moie fine fiaction piouuceu in compaieu with TBNY followeu the same
pattein as shown foi FFBN in Figuie 0n aveiage FFBN iepiesenteu anu of the TBNY in
anu iespectively A significant inteiaction P was also founu between
yeais anu cuts foi giowth iate which was at least lowei in Cuts anu in compaieu
with Theie was a significant P inteiaction between yeais anu cuts foi plant height
which was highei in in the fiist thiee Cuts while plant height at Cut was highei in
It is impoitant to highlight that most vaiiables geneially followeu the iainfall pattein showing a
uiastic inciease staiting in Nay eveiy yeai The cumulative BN yielu fiom the iainy season Cuts
was uouble that foi the uiy season Cuts
11


Fig 4 Effect of inteiactions between eight uiffeient cuts in two subsequent yeais on a fine fiaction uiy
mattei yielu b giowth iate anu c plant height Cuts C to C coiiesponu to the uiy season Novembei
Apiil anu cuts C to C to the iainy season Nay0ctobei
12

Figuie shows significant inteiactions between feitilization level anu cuts foi the vaiiables FFBN uR
anu B uue to minoi uiffeiences between N anu N at uiffeient Cuts

Fig 5 Effect of inteiactions between foui uiffeient levels of nitiogen feitilization kg N ha

yeai

anu eight
uiffeient cuts on a fine fiaction uiy mattei yielu b giowth iate anu c plant height Cuts C to C
coiiesponu to the uiy season NovembeiApiil anu cuts C to C to the iainy season Nay0ctobei
13

Bowevei on an oveiall level theie weie no uiffeiences between N anu N foi these vaiiables while the
levels N anu N consistently uiffeieu significantly fiom the highei feitilization levels Pon all
cutting occasions At Cut the levels N anu N piouuceu above ton ha

FFBN each while anu ton

ha

weie obtaineu on aveiage foi the same fiaction with levels N anu N iespectively A similai
pattein was obseiveu in giowth iate Plant height in uiffeient tieatments followeu moie oi less the
same tienu Bowevei level N uiffeieu fiom the othei tieatments by showing a uiastic ueciease in
height in Cut which was not obseiveu in the othei tieatments The gap between the two high levels of
feitilization anu the otheis was laigei anu moie eviuent uuiing the iainy season foi fine fiaction yielu
anu giowth iate
Chemical composition
Table shows the effect of the two uiffeient planting uensities on the chemical composition of Moringa
oleifera biomass at the cut with peak BN piouuction Cut No significant uiffeiences weie founu
between the planting uensities iegaiuing the chemical composition of the fiactions
Table Chemical composition of fine anu coaise fiactions of Moringa oleifera at two uiffeient planting
uensities in Nanagua Nicaiagua Least squaie means anu stanuaiu eiioi SE

vaiiable Planting uensity plant ha

SE
Fine fraction
Biy mattei g kg

g kg
1
DM*
Ciuue piotein
NBF
ABF
Lignin
Ash
Coarse fraction
Biy mattei g kg

g kg
1
DM*
Ciuue piotein
NBF
ABF
Lignin
Ash
BNuiy mattei

No significant uiffeiences weie founu between planting uensities foi the chemical composition paiameteis uesciibeu
in the table NBF neutial ueteigent fibie ABF aciu ueteigent fibie
The effect of feitilization on the chemical composition of uiffeient fiactions of Moringa oleifera biomass
is shown in Table In the fine fiaction significant uiffeiences P weie founu between N anu N
iegaiuing BN NBF anu lignin but no significant uiffeiences weie founu between feitilization
tieatments in CP ABF oi ash content
14

In the coaise fiaction no significant uiffeiences weie founu in NBF oi lignin content Bowevei N was
significantly lowei P in BN content compaieu with the othei feitilization levels Although no
significant uiffeiences weie founu between N anu N they weie both significantly highei P than
N in teims of CP concentiation Fuitheimoie significantly lowei ABF content P was founu in
the coaise fiaction of the N level compaieu with the othei feitilization levels
Table Chemical composition of fine anu coaise fiaction of Moringa oleifera biomass at foui uiffeient levels
of nitiogen feitilization N N N anu N kg N ha

yeai

Least squaie means anu

Stanuaiu eiioi SE
vaiiable Feitilization level SE
N N N N
Fine fraction
BN g kg

a

a

ab

b

g kg
1
DM
Ciuue piotein
NBF
b

ab

a

a

ABF
Lignin
b

ab

a

a

Ash
Coarse fraction
BN g kg

a

a

ab

b

g kg
1
DM
Ciuue piotein
b

ab

a

a

NBF
ABF
a

a

a

b

Lignin
Ash
ab

b

a

a

Neans within iows with uiffeient supeisciipts aie significantly uiffeient P In cases wheie SE values uiffeieu slightly between
tieatments the highest SE value is iepoiteu NBF neutial ueteigent fibie ABF aciu ueteigent fibie
Mortality
Buiing the fiist expeiimental yeai the moitality among plots was veiy low
Buiing the pieviously mentioneu attack by teimites gave a moitality of appioximately
which was a significant inciease P in moitality of compaieu with the pievious
yeai The uamage fiom the teimite attack was evenly uistiibuteu ovei tieatments anu no single
tieatment was moie affecteu than the otheis Bence no significant uiffeiences in moitality weie founu
between feitilization levels oi between planting uensities
15

Discussion
Effect of planting density on biomass production
A positive ielationship between planting uensity anu biomass yielu in tiopical tiee legumes such as
Calliandra calothyrsus Sesbania grandiflora Leucaena leucocephala anu Gliricidia sepium has been
iepoiteu Ella et al Although theie is uecieaseu yielu pei plant at highei uensities this ueciease
is moie than compensateu foi by incieasing yielu pei unit aiea Ella et al In Noiinga Foiul et al
iepoiteu a positive lineai ielationship between incieasing uensity anu biomass yielu wheieas
ReyesSnchez et al. b iepoiteu no significant uiffeiences in yielu between planting uensities of
anu plants ha

Nanh et al also iepoiteu no effect of uensity on

biomass yielu when uensities of plants ha

oi lowei weie useu Bowevei the iesults of the

piesent stuuy show that highei uensity gave highei yielu of Noiinga The TBNY obtaineu at the highei
uensity of plants ha

gave a highei yielu than the ton ha

iepoiteu by ReyesSnchez et
al b at plants ha

No inteiaction between yeai anu planting uensity was obseiveu in

the piesent stuuy uiowth iate was uouble at the highei planting uensity compaieu with the lowei
which is in complete accoiuance with the total uiy mattei yielu uiowth iate of Noiinga in the piesent
stuuy at the two planting uensities was ton ha

uay

which in geneial was consiueiably

highei than the anu ton ha

uay

iepoiteu foi the fouuei tiees Leucaena leucocephala

Shelton anu Biewbakei anu Gliricidia sepium Simons anu Stewait iespectively
Effect of nitrogen fertilization on biomass production
In cut anu caiiy systems huge amounts of nutiients aie iemoveu fiom the soil wheie the ciop is
haivesteu In a peiennial ciop such as Noiinga this leaus to a ieuuction in biomass yielus ovei time
especially if high planting uensities aie useu A ieuuction of about in biomass yielu was iepoiteu
by ReyesSnchez et al b foi unfeitilizeu Noiinga in a twoyeai tiial The way to keep oi
inciease yielu in peiennial ciops is to maintain the soil nutiients at a sufficiently high level by
feitilization Nitiogen plays a pivotal iole in plant uevelopment anu biomass yielu anu was theiefoie
the focus of this expeiiment 0thei stuuies have iepoiteu the positive effect of N feitilization on
Noiinga yielu Bash anu uupta }yothi anu Babu Pamo et al. Bowevei those stuuies
weie peifoimeu at laboiatoiy level anu the iesults cannot be uiiectly extiapolateu to fielu conuitions
This stuuy showeu an inteiaction between level of feitilization anu yeai The FFBN anu uR weie
similai but weie slightly highei in the fiist yeai compaieu with the seconu yeai at feitilization levels N
16

anu N These iesults coulu be uue to the combineu effect of a uiiei enviionment anu low soil N
concentiation which stunteu those vaiiables in the seconu yeai Fuitheimoie the effects of low N
feitilization aie likely to become moie pionounceu ovei time as the soil giauually becomes uepleteu of
N thiough the iegiowth being continuously haivesteu eveiy uays Biy mattei piouuction incieaseu
annually when N feitilization incieaseu N anu N but ueclineu in N anu N The iesponse of plant
giowth to N application is wiuely iecognizeu in conuitions wheie N is a limiting factoi Salmeion
Niianua et al The iesponse obseiveu heie in BN yielu shows that N feitilization was a key
factoi foi biomass piouuction Bowevei futuie soil analyses shoulu not only ueteimine the ability to
supply the N iequiiements of plants anu animals but also the complexity of soil biological piopeities
Abbot anu Nuiphy Biinkwatei et al Theie weie thiee impoitant exteinal factois besiues
the expeiimental vaiiables which affecteu biomass piouuction in this expeiiment The fiist was soil N
content as uiscusseu above the seconu was the amount of iainfall which was highei uuiing
than in anu the thiiu factoi conceineu the plant itself in that newly establisheu
Noiinga uevelopeu only one tiunk which was obviously thickei but as soon as the piuning staiteu the
iegiowth was thinnei This confiims finuings by ReyesSnchez et al b that the coaise fiaction
of Noiinga was ieuuceu by in the seconu yeai compaieu with the fiist Regaiuless of yeai in this
stuuy no uiffeiences in TBNY weie founu between N anu N This lack of uiffeience can be explaineu
by the Nitscheilich law of uiminishing ietuins anu also suggests that the optimum nutiient
iequiiement foi Noiinga is aiounu kg N ha

yeai

N
While the inteiaction between yeais anu cuts was not veiy cleai uuiing the uiy season Cuts to
the uiiei seconu yeai piouuceu laigei yielus with iegaiu to FFBN compaieu with the
fiist yeai Plants tenu to giow thinnei in uiy conuitions compaieu with wet wheie plants
not only giow tallei but also thickei with a highei coaise fiaction Nommei et al In geneial
highei levels of feitilization geneiateu laigei yielus in the iainy season Bowevei while no uiffeiences
coulu be founu between N anu N in teims of FFBN uR anu B uuiing all cuts significant uiffeiences
P between these levels compaieu with the othei two N levels weie iepeateuly founu in all cuts
uuiing the annual cycle Conveisely significant uiffeiences P in FFBN uR anu B weie founu
between N anu N uuiing the iainy season Besiues the obvious inciease in watei availability uuiing
the iainy season anu its effect on biomass piouuction the feitilization stiategy shoulu be taken into
account Feitilization was peifoimeu in }une anu 0ctobei anu unuei wet conuitions uiea is usually
iapiuly conveiteu to ammonium anu can be absoibeu by plants eithei in ammonium foim oi as nitiate
Theiefoie the N supplieu heie was fully available to the plants uuiing that peiiou of time as iepoiteu
by Tannei et al Anothei impoitant point is the effect of soil pB on N uptake which is at its
optimum aiounu neutial pB
17

Effect of planting density on chemical composition of Moringa
Planting uensity hau no significant effect on the chemical composition of Noiinga uuiing the fiist oi
seconu yeai No effect on chemical composition of foliage is expecteu when planting uensities aie not
high anu theiefoie the competition among plant foi nutiients is not ciitical Similai iesults weie
iepoiteu by Nanh et al anu ReyesSnchez et al b Stuuies on othei tiees such as Morus
alba anu shiubs such as Manihot sculenta also iepoiteu no uiiect effect of planting uensity on CP
concentiation of biomass ventuia anu Pulgai Boschini et al Ciuue piotein contents
iepoiteu in this stuuy foi the fine fiaction weie within the g CP kg

BN iange iepoiteu by
ReyesSnchez et al b anu Soliva et al
Effect of N fertilization on chemical composition of Moringa
Theie weie no obvious effects of N feitilization level on CP ABF anu ash content of the fine fiaction
The effect of N feitilization on chemical composition of foliage seems to vaiy among plant species anu
to the best of oui knowleuge no othei iepoits on this in Noiinga aie available foi compaiison When
the shiub Manihot sculenta was stuuieu a significant effect of feitilization on CP NBF ABF anu ash
content was iepoiteu Phengvichith et al but with the tiee Morus spp Rouiiguez et al
iepoiteu no such effect Biffeient iesponse to N feitilization can also be seen among tiee species
Baitley et al founu significant uiffeiences in N content on foliage of Picea sitchensis uepenuing
on feitilization but no effect on chemical composition of Calluna vulgaris foliage in the same
expeiiment The uiffeiences in BN content in the fine fiaction of Noiinga obseiveu in the piesent
stuuy can be explaineu by the plants giowth pattein as plants unuei high N levels N anu N tenueu
to giowth tallei but thinnei anu weie moie succulent than plants unuei N anu N This was confiimeu
by the uiastic inciease in the fine fiaction at the highei levels of N feitilization compaieu with the
lowei levels Bowevei tallei plants neeu moie stiuctuial components to maintain stability which may
explain the slightly highei NBF content in FFBN foi level N compaieu with N
As expecteu the content of BN NBF ABF anu lignin in the coaise fiaction was highei than that in the
fine fiaction Bowevei the coaise fiaction with levels N anu N pioveu to be moie succulent than the
coaise fiaction of the lowei N levels as inuicateu by the lowei BN content anu highei CP
concentiation as a iesult of a highei availability of N to the plants
18

Mortality
Even though Noiinga is sometimes claimeu to be pestiesistant Palaua anu Chang it is
impoitant to be awaie that it can be infesteu by pests just like any othei ciop In this stuuy a minoi
attack by ants Formica spp was obseiveu when the ciop staiteu to spiout but no majoi uamage was
noteu Bowevei an infestation of teimites was obseiveu uuiing anu moitality in
that yeai was the iesult of uamage fiom teimite attack Naitin anu Rubeit cautioneu against
the use of Noiinga in Pueito Rico uue to its high susceptibility to teimites while }iang anu Liu
classifieu teimites among the five main pests of Noiinga 0nce pest contiol was caiiieu out in the
piesent stuuy no fuithei uamage was founu anu peicentage moitality was at veiy low levels
Conclusions
This stuuy showeu that unuei tiopical uiy foiest conuitions anu when P anu K aie available in the soil
Noiinga can maintain high biomass yielu ovei time pioviueu that N is supplieu in sufficient amounts
to covei what is iemoveu at haivest In the piesent stuuy a combination of plants ha

anu
kg N ha

yeai

was best uue to the possibility of achieving veiy high total uiy mattei yielu with a high
piopoition of fine fiaction Bowevei expeiiments ovei longei time peiious such as five yeais woulu be
valuable to fuithei veiify that stable sustainable piouuction of appioximately tons ha

can be
maintaineu at this planting uensity anu feitilizei level Noiinga can help small anu meuiumscale
faimeis oveicome shoitages of goou quality feeus anu theiefoie to sustain anu impiove theii livestock
systems
Acknowledgments
The funuing foi this ieseaich pioviueu by the Sweuish Inteinational Bevelopment Agency SIBA is
giatefully acknowleugeu Bi Lestei Rocha anu }ohannes Foikman aie also sinceiely acknowleugeu foi
theii help anu suppoit with the statistical analysis in this stuuy

References
Abbot L Nuiphy B What is soil biological feitility In Abbot L Nuiphy B eus Soil
biological feitility Spiingei The Netheilanus
19

Aganga A Tshwenyane S Feeuing values anu antinutiitive factois of foiage tiee legumes Pak }
of Nuti
Anwai F Latif S Ashiaf N uilani A Moringa oleifera A foou plant with multiple meuicinal
uses Phytothei Res
A0AC 0fficial methous of analysis
th
euition Association of 0fficial Analytical Chemists
Washington 0S
A0AC 0fficial methous of analysis
th
euition Association of 0fficial Analytical Chemists
uaitheisbuig 0S
Boschini C Boimonu B Castio A Composicion quimica ue la moieia Morus alba paia uso en
la alimentacion animal uensiuaues y fiecuencias ue poua Agion Nesoam
Calub B Evaluation of inuigenous anu natuializeu multipuipose tiees as alteinative to Leucaena
leucocephala in hillylanu faiming SeminaiWoikshop Towaius moie effective utilization of iesouices
foi sustaineu uevelopment Nakati Netio Nanila Philippines Philippine Techn }
Bash S uupta N Effect of inoiganic oiganic anu bio feitilizei on giowth of hybiiu Moringa
oleifera PKN Sci
Biinkwatei L Schipanski N Snapp S }ackson L Ecological baseu nutiient management In
Snapp S Pounu B Agiicultuial System 0SA
Buke } Banubook of eneigy ciops 0puateu }anuaiy th Retiieveu 0ctobei th fiom
httpwwwhoitpuiuueeuunewciopuukeeneigyNoiingaoleifeiahtml
Ella A }acobsen C Sti W Blaii u Effect of plant uensity anu cutting fiequency on the
piouuctivity of foui tiee legumes Tiop uiassl
FA0 FA00nesco Soil Nap of the Woilu Reviseu legenu with coiiections anu upuates Woilu
Soil Resouices Repoit FA0 Rome
Feiieiia P Faiia B 0liveiia } Caivalho A Moringa oleifera bioactive compounus anu
nutiitional potential Rev ue Nutiio
Foiul N Nakkai B Beckei K The potential of Moringa oleifera foi agiicultuial anu inuustiial
uses What uevelopment potential foi Noiinga piouucts Bai Es Salaam
20

Baitley S Nelson K uoiman N The effect of feitilisei anu shauing on plant chemical
composition anu palatability to 0ikney voles Microtus arvalis orcadensis 0ikos
Bolmann F Lascano C eus Feeuing systems with foiage legume to intensify uaiiy piouuction
in Latin Ameiica anu the Caiibbean A pioject executeu by the Tiopileche Consoitium CIAT Centio
Inteinacional ue Agiicultuia Tiopical anu ILRI Inteinational Livestock Reseaich Institute Naiiobi
Kenya pp
INETER Instituto Nicaiaguense ue Estuuios Teiiitoiiales Boletin climtico Nanagua
Nicaiagua Biieccion geneial ue meteoiologia
}iang u Liu C Nain uiseases anu pests on Noiinga tiees in Xishaungbanna Tiop Agiic Sci anu
Tech
}yothi u Babu R uiaueu uoses of nitiogen on uiumstick Moringa pterygosperma uoeitn vai
PKN }ouinal of ieseaich ANuRA0
Leng R Tiee legumes in iuminant nutiition FA0 Animal piouuction anu health Papei FA0
Rome Italy
Nakkai B Beckei K Nutiitional value anu antinutiitional components of whole anu ethanol
extiacteu Moringa oleifera leaves Anim Feeu Sci anu Tech
Nanh L Nguyen N Ngoi T Intiouuction anu evaluation of Moringa oleifera foi biomass
piouuction anu as feeu foi goats in the Nekong uelta Livest Res foi Ruial Bev
Naitin F Rubeit R Euible leaves of the tiopics 0SBA Nayaguez Pueito Rico
Nommei L Lenssen } Bubei B vissei E Be Kioon B. Ecophysiological ueteiminants of plant
peifoimance unuei floouing a compaiative stuuy among seven plant families. }ouinal of Ecology .
.
Noiton } The hoiseiauish tiee Moringa pterygosperma NoiingaceaeA boon to aiiu lanus
Econ Bot
0liveiia } Silveiia S vasconcelos I Cavaua B Noieiia R Compositional anu nutiitional
attiibutes of seeus fiom the multiple puipose tiee Moringa oleifera Lamaick } of the Scie of Foou anu
Agiic
21

0liveiia } Souto } Santos R Souto P Naioi }nioi S Auubao com uifeientes esteicos no
cultivo ue moiinga Moringa oleifera LAN Rev veiue ue Agioecol e Besenvolvimento Sustentvel.
Palaua N Chang L Suggesteu cultuial piactices foi Noiinga Asian vegetable Reseaich anu
Bevelopment Centei Pub pp
Pamo E Boukila B Tonfack L Nomo N Kana } Tenuonkeng F Influence ue la fumuie
oiganique uu NPK et uu mlange ues ueux feitilisants sui la cioissance ue Moringa oleifera Lam uans
l0uest Cameioun LivestRes foi Ruial Bev
Phengvichit v Leuin S Boine P Leuin I Effects of uiffeient feitiliseis anu haivest fiequencies
on foliage anu tubei yielu anu chemical composition of foliage fiom two cassava Manihot esculenta
Ciantz vaiieties Tiop anu Subtiop Agioecosyst
Ramachanuian C Petei K uopalakiishnan P Biumstick Moringa oleifera A multipuipose
Inuian vegetable Econ Bot
ReyesSnchez N Spoinuly E Leuin I a Effect of feeuing uiffeient levels of foliage of Moringa
oleifera to cieole uaiiy cows on intake uigestibility milk piouuction anu composition Livest Sci

ReyesSnchez N Leuin S Leuin I b Biomass piouuction anu chemical composition of Moringa
oleifera unuei uiffeient management iegimes in Nicaiagua Agiofoiestiy Syst
Rouiiguez C Aiias R Quiones } Efecto ue la fiecuencia ue poua y el nivel ue feitilizacion
nitiogenaua sobie el ienuimiento y caliuau ue la biomasa ue Noieia Morus spp en el tiopico seco ue
uuatemala In Benaviuez } eu Aiboles y aibustos foiiajeios en Amiica cential vol Seiie
tcnica CATIE pp
SalmeionNiianua F Bath B Eckeisten B Foikman } Wisvtau N Abovegiounu nitiogen in
ielation to estimateu total plant uptake in maize anu bean Nutiient Cycling Agioecosyst
SAS Statistical analysis system institute Inc 0sei guiue veision 0S
Shelton B Biewbakei } Leucaena leucocephala The most wiuely useu foiage tiee legume In
uutteiiuge R Shelton B eus Foiage tiee legumes in tiopical agiicultuie Tiop uiassl Soc of
Austialia Inc Queenslanu
22

Simons A Stewait } Gliricidia sepium a Nultipuipose foiage tiee legume In uutteiiuge R
Shelton B eus Foiage tiee legumes in tiopical agiicultuie Tiop uiassl Soc of Austialia Inc
Queenslanu
Smith 0 0sing fouuei fiom tiees anu shiubs to feeu livestock in the tiopics FA0 Rome Italy
Soliva C Kieuzei N Foiul N Foiul u Nachmllei A Bess B Feeuing value of whole anu
extiacteu Moringa oleifera leaves foi iuminants anu theii effects on iuminal feimentation in vitio
Anim Feeu Sci anu Tech
Szott L Kass B Feitilizeis in agiofoiestiy systems Agiofoiestiy Syst
Tannei } Kapos v Fieskos S Bealey } Theobalu A Nitiogen anu phosphoius feitilization of
}amaican montane foiest tiees }ouinal of Tiopical Ecology B0I S
Toleuo } ShultzeKiaft R Netouologia paia la evaluacion agionomica ue pastos tiopicales In
Toleuo } eu Nanual paia la evaluacion agionomicaReu inteinacional ue evaluacion ue pastos
tiopicales Centio Inteinacional ue Agiicultuia Tiopical pp
0SBA 0niteu States Bepaitment of Agiicultuie Soil suivey laboiatoiy infoimation manual
Lincoln Nebiaska pp
van Soest P Robeitson Lewis B Nethous foi uietaiy fibei neutialueteigent fibei anu non
staich polysacchaiiues in ielation to animal nutiition } of Baiiy Sci
ventuia } Pulgai R Efecto ue la uensiuau ue siembia y fiecuencia ue coite sobie los
components ue la piouuccion y follaje ue yucca Manihot esculenta Ciantz Revue Agionom
Zamoia Sh uaicia } Bonilla u Aguilai B Baivey C Ibiahim N 0se of tiee fiuits anu foliage
to feeu cattle in the uiy season Boaco Nicaiagua Agiofoi en las Am

I
Moringa (Moringa oleifera) leaf meal as a source of protein in locally
produced concentrates for dairy cows fed low protein diets in tropical areas
B. Mendieta-Araica
b
, R. Sprndly
a
, N. Reyes-Snchez
b
, E. Sprndly
a,

a
Department of Animal Nutrition and Management, Swedish University of Agricultural Sciences, P.O. Box 7024, SE-750 07, Uppsala, Sweden
b
Facultad de Ciencia Animal, Universidad Nacional Agraria, P.O. Box 453, Managua, Nicaragua
a r t i c l e i n f o a b s t r a c t
Article history:
Received 22 June 2010
Received in revised form 22 September 2010
Accepted 22 September 2010
The effect on milk yield, milk composition and ration digestibility of using Moringa leaf meal as
a protein source in concentrate given to six lactating dairy cows fed a basal Elephant grass diet
was tested using a changeover 33 Latin square design, replicated twice. The basal Elephant
grass diet and a concentrate containing 20% soybean meal was compared with a concentrate
where the soybean meal was replaced with the same amount of Moringa leaf meal. In the third
diet commercially available components were used to compose an Iso concentrate with the
same energy and protein content as the concentrate containing Moringa leaf meal. The intake
of dry matter, organic matter, neutral detergent bre and acid detergent bre did not differ
signicantly between treatments and averaged 15.4, 13.9, 7.2 and 5.9 kg day
1
, respectively,
while crude protein (CP) intake was higher (Pb0.001) for the soybean meal treatment
compared to the other treatments, 1.7 and 1.2 kg CP day
1
, respectively. The treatments did
not differ with regard to digestibility with the exception of CP digestibility, which was
signicantly higher in the soybean meal treatment compared with the Iso concentrate, 0.70 and
0.62, respectively. Mean daily milk yield was signicantly higher (Pb0.05) when cows were
given soybean meal compared with both Moringa leaf meal and the optimized concentrate,
13.2, 12.3 and 12.1 kg day
1
, respectively. There was no signicant difference between
treatments in either the milk composition, or the organoleptic characteristics of the milk. The
conclusion is that locally produced Moringa leaf meal can, at the same protein and energy
levels, successfully replace the commercial constituents in concentrate for dairy cows.
2010 Elsevier B.V. All rights reserved.
Keywords:
Moringa leaf meal
Milk yield
Milk composition
Organoleptic characteristics
1. Introduction
One of the most important constraints in tropical livestock
production systems is underfeeding due to limitations in both
quantity and quality of feed. This has been recognized by
many authors (Franziska and Baccini, 2005; Olafadehan and
Adewumi, 2009, 2010). These restrictions lead to low milk
yields or growth rates which, in turn, gives low net incomes
for farmers (Olafadehan and Adewumi, 2008; de Leeuw et al.,
1999). This is particularly pronounced during the dry season,
when natural pastures are mature and dry, and therefore
have a low nutritive value. In many areas of the tropics,
Elephant grass (Pennisetum purpureum) constitutes the basal
diet for dairy cows (Sarwatt et al., 2004; Shem et al., 2003).
However, due to the relatively lowquality of Elephant grass, it
is essential to provide a protein-rich feed supplement.
Dairy production in tropical areas is a complex systemand
cannot be seen isolated from the economical and social
dimension in which the farmers live. Supplementation with
conventional concentrates during the dry season is generally
too costly and the levels of concentrate feeding are therefore
low (de Leeuw et al., 1999). The use of concentrate or other
by-products as feed supplements to dairy cows on small
farms will depend on the access to cash and the price of the
feeds and also the cost of transportation and availability.
Therefore there is a need to nd alternative low-cost
Livestock Science 137 (2011) 1017
Corresponding author. Kungsngen Research Centre, SE-753 23 Uppsala,
Sweden. Tel.: +46 18 671632; fax: +46 18 672948.
E-mail address: eva.sporndly@huv.slu.se (E. Sprndly).
1871-1413/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.livsci.2010.09.021
Contents lists available at ScienceDirect
Livestock Science
j our nal homepage: www. el sevi er. com/ l ocat e/ l i vsci
supplements which can be cultivated by the dairy farmer and
are available all-year around. This will allow farmers to
improve the nutritional level in dairy production in the
tropics and step by step improve the economy in small scale
dairy production.
An important concentrate ingredient in many countries is
soybean meal (SBM) which has a high crude protein (CP)
content varying from 437 to 480 g kg
1
dry matter (DM)
(Broderick et al., 1990; Castillo et al., 2001) and provides a
combination of amino acids that can support high milk yields
(Broderick et al., 1990; McDonald et al., 2002). However, local
conditions for soybean cultivation are not always favourable
and in such areas it is expensive to import SBM. In tropical
countries such as Nicaragua, SBMis one of the most expensive
feed ingredients in concentrate feeds. During the last four
years, the price of SBM has increased almost 200% (USDA,
2009) while the price of other protein-rich feeds such as
peanut meal has increased only 9% (MAGFOR, 2008). In some
tropical regions urea has been used as an alternative nitrogen
source in ruminant diets but in many countries such as
Nicaragua, urea is an imported market commodity with
variable availability and price. Small farmers are reluctant to
use urea because they need to purchase it on the market,
availability is uncertain and it is considered difcult to use as
a feed. Faced with these facts, it would be benecial for small
farmers in to replace all or part of the SBM in their
concentrate mixture with a more economical protein source
that could be grown locally with limited resources beside
labour.
Leaf meal is a good and cheap source of protein
(Duckworth and Woodham, 1961; Paterson et al., 1998).
Different forage trees and shrubs, such as Chromolaena
odorata (Fasuyi et al., 2005), Leucaena leucocephala
(Kakengi et al., 2001), Morus alba, Azadirachta indica
(Patra et al., 2003) and Acacia karroo (Mapiye et al.,
2009) have been fed to goats, layers, steers and dairy cows,
with good production results. However, the presence of
various anti-nutritional compounds in foliage from trees
and their deleterious effects in animals has also been
discussed (Ghosh et al., 2007; Hammond, 1995). Largely
due to the presence of anti-nutritional compounds such as
mimosine, cyanogenic glycosides, condensed tannins and
alkaloids, the use of forage trees and shrubs has been
limited and ad libitum feeding of these forages is rarely
used in livestock feeding.
Another potential protein source for livestock production
is Moringa (Moringa oleifera). Although it is a widespread,
drought tolerant tree with a high DM yield in the tropics
(Reyes-Snchez et al., 2006), the potential for using Moringa
as animal feed is still underappreciated. It is a tree with a high
CP content, varying from 179 to 268 g kg DM
1
(Reyes-
Snchez et al., 2006; Mendieta-Araica et al., 2009), and with
negligible amounts of tannins, trypsin and amylase inhibitors
(Becker, 1995; Gidamis et al., 2003; Makkar and Becker,
1997). Moringa has also been reported to be a valuable
component in human food due to its adequate amino acid
prole and CP content, its high level of vitamin A and its low
level of anti-nutritional compounds (Anhwange et al., 2004;
Snchez-Machado et al., 2009). Due to this recent interest in
Moringa, feeding trials using fresh Moringa have been
performed with many types of animals such as pigs, goats
and creole cows (Ly et al., 2001; Aregheore, 2002; Reyes-
Snchez et al., 2006). Feeding fresh Moringa is convenient but
there is a large variation in production over the year.
Therefore, Moringa leaf meal (MLM) is an interesting product
as it can be produced during periods of high yields and later
used for feeding during the dry season when high quality feed
resources are scarce. An important advantage with the
production of Moringa leaf meal (MLM) is that the required
technology is affordable and feasible even for small farmers.
Moringa foliage (branches, twigs and leaves) can be obtained
either from pure crop plots or live fences, cut with machete
and sun-dried on a black plastic sheet placed on the ground
(Olsson and Wilgert, 2007). Based on the authors' experience,
the whole drying process can be completed in 72 h rendering
approximately 1 kg of MLM from 10 kg fresh material. After
drying, the leaves can be removed by simple threshing and
the remaining small dry leaves can be crushed or ground by
hand to obtain MLM.
In recent years, the interest in MLM as a diet component
in animal production has received some attention by
researchers who have reported promising production
results in sh (Richter et al., 2003), sheep (Murro et al.,
2003) and laying hens (Kakengi et al., 2007). Furthermore,
in an interesting experiment performed with cross-bred
dairy cows MLM was compared with cotton seed cake
(CSC) as a concentrate component together with maize
bran and minerals (Sarwatt et al., 2004). The cows were fed
a basal Elephant grass diet together with one of three
concentrate mixtures. Moringa leaf meal substituted 43, 73
and 100% of the CSC in these mixtures. The cows that were
fed higher proportions of MLM yielded signicantly more
milk indicating that MLM is an interesting feed resource in
dairy cow diets.
However, the number of dairy cow experiments with
MLM is limited (Sarwatt et al., 2004) and it is therefore
interesting to study the potential of MLM as an alternative
protein source for milk production further, and explore the
labour requirements for small scale on-farm production of
MLM. Therefore, the aim of this study was to evaluate how
MLM compares to commercial concentrate constituents with
regard to milk yield, milk composition and ration digestibility
and to estimate the work needed to produce MLM under
small scale farming conditions.
2. Materials and methods
2.1. Location
The experiment was carried out during the dry season at
Santa Ana Farm in Masaya, Nicaragua, located at 132916.5
N and 605510 W. The average annual temperature in
Masaya is 26.6 C and the mean annual rainfall is 1361.3 mm,
with a marked dry season (NovemberMay).
2.2. Treatments, experimental design and management
A basal Elephant grass diet with a concentrate containing
20% SBM was tested against the same basal diet with a
concentrate where the SBM was replaced with the same
amount of MLM on weight basis. In the third experimental
diet, commercially available components (including SBM)
11 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
were used to compose an Iso concentrate with the same
energy and protein content as the concentrate containing
MLM.
The concentrate mixtures used in the three experimental
treatments are presented in Table 1. The SBMconcentrate had
the same composition as concentrates commonly used in the
dairy industry of Central America. In the MLM concentrate,
the SBM was replaced by MLM on weight basis. The Iso
concentrate had the same energy and protein content as the
MLM concentrate but contained the cheapest available
commercially ingredients instead.
The experiment was designed as a Changeover 33
Latin Square, as described by Patterson and Lucas (1962),
replicated in two orthogonal Latin squares. Each experi-
mental period consisted of 2 weeks for treatment adapta-
tion and 2 weeks of data collection with regard to milk
yield and feed intake. The last week of each period was
used to estimate digestibility.
Six dairy cows from the farm herd, in their second or third
lactation and weighing 46722 kg were used in the trial. The
cows were in their fourth week of lactation at the start of the
experiment. All of the cows were treated according to EC
Directive 86/609/EEC for animal experiments. The animals
were loose conned in individual, well-ventilated stalls with
concrete oor which was covered with rubber mats during
faecal collection periods. Before the start of the trial, the cows
were injected with Vitamin A (625,000 IU), Vitamin D
3
(125,000 IU) and Vitamin E (125 IU); treated against external
and internal parasites and vaccinated against anthrax. Water
was provided ad libitum and the cows had access to a
commercial mineral supplement with Ca, P, Mg and trace
elements.
The feed allotment was planned to full the protein
requirement (NRC, 2001) of CP for the SBM diet when
adopting the DM intake of 3.2 of body weight recom-
mended in the region (Hazard, 1990; Romero and Gonzlez,
2004). Sixty percent of the expected DM intake was given
as roughage in the form of Elephant grass and the
remaining 40% was given as one of the three concentrates.
These proportions represent approximately those usually
used by dairy farmers in the Central American region
(Castro Ramrez, 2002; Vlez, 1997). Roughages were
offered individually in separate feed troughs twice per
day, at 07:00 h and 17:00 h. The concentrates were fed
individually during milking, at 05:00 h and 16:00 h. The DM
content of Elephant grass was determined twice per week
using a microwave oven according to the procedure
described by Undersander et al. (1993).
The offered amounts of feed and occasional refusals were
weighed daily and sampling of feeds was performed as
follows. One kilogram of offered roughage per cow per day
was collected and immediately frozen at 18 C. After every
period the frozen samples of offered feed for each cow were
thawed and pooled into one sample per period for chemical
analysis. The occasional refused feed was individually
sampled and frozen immediately at 18 C. At the end of
each experimental period these individual samples of refusals
were thawed and sent to a laboratory for chemical analysis.
From each concentrate in the experiment 1 kg from every
delivered batch was collected for analysis, giving a total of
three samples per concentrate corresponding to the three
periods in the experiment.
The cows were hand-milked and at each milking the yield
was weighed and recorded. From each cow and milking a
sample of 100 ml was collected and immediately refrigerated
at 4 C. Milk samples from each cow were pooled into one
sample per week during the data collection period and then
analyzed for fat and protein content, the same milk sample
procedure was used for the organoleptic test.
2.3. Feed preparation
The branches with leaves and soft twigs used for the
production of MLM were collected from M. oleifera trees in
the experimental area by cutting every 45 days. They were
then sun-dried for 24 h before the partially dried leaves were
removed by threshing and then sun-dried again approxi-
mately 48 h on black plastic sheets. The dried leaves were
nely ground in a hammer mill, packed in sacks and stored in
a well-ventilated storeroom. During the production of MLM
for the experiment, the amount of fresh material harvested
and the amount of MLM produced was weighed. Further-
more, the work used for harvesting Moringa and the amount
of work used in the drying process was registered.
The other concentrate ingredients were purchased in the
local market and the experimental concentrate mixtures
(Table 1) were produced at the Feed Concentrate Plant of the
National Agrarian University, Nicaragua. Before the start of
the experiment, the eld with Elephant grass (P. purpureum)
was divided into several plots that were harvested at regular
intervals before experimental start. This was done to enable
regrowth cuts at similar intervals with the objective to
harvest at approximately 45 days regrowth throughout the
experiment. During the experiment, the elephant grass was
harvested daily at 463 days regrowth and offered fresh to
the cows, chopped into pieces of approximately 2 cm using a
tractor mounted forage chopper.
2.4. Digestibility study
Every day during the last week of each period, all the
faeces from each cow were collected manually. During the
collection period cows were supervised 24 h daily and any
time when a cow adopted the defecation position a shovel
was put under her tail to collect the faeces and avoid
contamination from urine and dirt from the stall oor.
Table 1
Proportions of feedstuffs in % wet weight in the concentrate mixtures.
Ingredients MLM
concentrate
Iso
concentrate
SBM
concentrate
Sorghum 30.1 55.3 30.1
Rice polishing 41.0 25.1 41.0
Sugar cane
molasses
5.9 5.9 5.9
Peanut meal 1.0 9.4 1.0
Calcium carbonate 1.5 1.5 1.5
Salt 0.5 0.5 0.5
Moringa leaf meal 20.0 0.0 0.0
Soybean meal 0.0 2.3 20.0
MLM: moringa leaf meal, SBM: soybean meal.
12 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
Thereafter, the faecal material was put into a large,
individually marked container for the cow and covered
with a lid to avoid evaporation. Once daily, the faecal
contents in the containers of each cow were weighed and
thoroughly mixed. Five percent of the daily faecal contents
from each cow was then taken as a subsample and frozen.
When the collection was complete, the subsamples were
thawed and mixed together into one homogeneous sample
per cow and period. Approximately 300 g of the mixture
from each animal was then taken as a faecal sample. The
results of the chemical analysis of faeces and feed were
used together with intake data to estimate apparent
digestibility.
2.5. Chemical analysis
The samples of offered feed, refused feed and faeces were
dried at 60 C and ground through a 1 mm sieve before
analysis. Ash and DM were analyzed using AOAC (1990)
procedures. The CP content was determined using the
Kjeldahl method (AOAC, 1984). Neutral detergent bre and
ADF were analyzed as described by Van Soest et al. (1991).
The apparent digestibility coefcient for DM was calculated
by comparing the dietary intake of constituents and the
amounts recovered in faeces. Nitrogen content in the milk
was determined using the Kjeldahl method and milk protein
content was calculated as N6.38. Milk fat was determined
using the Babcock method (Pereira, 1988), while total solids
and casein were analyzed according to AOAC (1984)
procedures. An organoleptic evaluation of the milk was
performed by an experienced panel of 15 persons. A triangle
difference test (Witting de Penna, 1981) was applied using a
milk sample with normal sensory characteristics (colour,
smell and taste) as standard.
Organic matter digestibility and the metabolisable energy
(ME) of the elephant grass were determined by in vitro
incubation in rumen liquid: 0.5 g dried sample mixed with
49 ml buffer and 1 ml rumen uid was incubated 96 h at
38 C. The residues were combusted to get the digestibility
coefcient of the organic matter and ME was then estimated
using equation presented by Lindgren (1979). For concen-
trates, ME was calculated based on the Weende analysis, as
described by McDonald et al. (1988).
2.6. Statistical analysis
The data was analyzed using the GLM procedure in the
SAS Version 9.1.2 (SAS, 2004). Tukey's pairwise comparison
procedure was used whenever the overall F-test of treatment
means showed a signicant result. The mathematical model
used was
Yjk =+Pk +Cj +Tl +jk, with j =1,,6 and k=1,2,3,
where was the overall mean, Pk the xed effect of
period, Cj the random effect of cow, Tl the xed effect of
treatment and jk the random residual error.
Carry-over effects from previous periods, the interaction
between periods and treatments were tested initially, as
described by Patterson and Lucas (1962), but were excluded
from the nal model because of lack of signicance (PN0.10).
3. Results
3.1. Feed intake and apparent digestibility
The MLM used in this experiment contained 292 g CP,
161 g neutral detergent bre (NDF), 151 g acid detergent
bre (ADF), 68 g lignin and 94 g ash per kg DM. The chemical
composition of the feeds presented in Table 2 shows that the
basal Elephant grass diet had a low CP content (34 g kg
1
DM) and a high bre content. The CP content in the SBM
concentrate was typical for commercial concentrates in
Central America.
The feed intake and apparent digestibility coefcients are
shown in Table 3. There were no signicant differences
among treatments with regard to feed intake or the intake
and digestibility of DM, organic matter, NDF and ADF
(Table 3). The CP and ME intake was higher for cows fed
the SBM concentrate when compared with the other treat-
ments. Compared to the NRC requirements (2001) the CP
intake covered 104%, 76% and 73% of requirements for the
SBM, MLM and Iso diets, respectively. The CP content of the
diet was 102 g kg
1
DM for cows fed SBM concentrate and
79 g kg
1
DMfor the other treatments. The starch intake was
also higher for cows fed the Iso concentrate. Furthermore, the
digestibility of CP was higher when cows were fed SBM
concentrate compared with Iso concentrate.
3.2. Milk yield and composition
The average daily milk yield during the experiment was
11.8 kg energy correctedmilk (ECM). Bothmeandaily milk and
ECMyield were signicantly (Pb0.05) higher when cows were
fed SBM concentrate compared with the other treatments
(Table 4). However, there was no signicant difference in milk
composition between treatments and, in average, the milk
contained 34.9 g kg
1
fat, 34.5 g kg
1
protein, 126.1 g kg
1
total solids and 27.4 g kg
1
DM casein. The hypothesis that
MLM would inuence the organoleptic characteristics of the
milk proved to be wrong: the colour, smell and taste of milk
from all treatments were classied as normal.
Table 2
Chemical composition of the concentrates and basal diet used in the
experiment, means and standard deviation (in parenthesis) n=3.
Nutrients Feeds
Elephant
grass
MLM
concentrate
Iso
concentrate
SBM
concentrate
DM, g kg
1
173 (33) 837 (12) 836 (15) 858 (14)
g kg
1
DM
Crude protein 34 (8) 154 (5) 153 (13) 221 (5)
Neutral detergent
bre
673 (52) 129 (11) 119 (8) 106 (7)
Acid detergent
bre
558 (23) 86 (2) 84 (8) 72 (0)
Lignin 105 (18) 28 (1) 25 (2) 21 (1)
Ash 119 (10) 91 (4) 76 (8) 82 (2)
Starch nd 337 (21) 462 (18) 354 (16)
G+F nd 25 (2) 16 (4) 14 (2)
ME

MJkg
1
DM 7.1 (0.4) 12.9 (0.3) 12.9 (0.2) 13.4 (0.2)
DM: dry matter, MLM: moringa leaf meal, SBM: soybean meal, G+F: glucose+
fructose, ME: metabolizable energy, * calculated, nd: not determined.
13 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
3.3. Labour requirements for Moringa leaf meal production
Weighing of the freshly harvested Moringa and the dried
leaves showed that 1 ton of fresh Moringa gave 123.5 kg of
dried Moringa leaves. At a planting density of 100,000
170,000 Moringa plants per ha and a harvesting interval of
45 days, the work required for 1 person to harvest 1 ton of
Moringa was 2.4 days. The work involved in spreading out
this amount of harvested Moringa, for threshing and the nal
drying on plastic sheets was 0.8 days for 1 man. The leaves
were, as mentioned earlier, ground at a feed plant and no
record of the work needed for manual grinding or crushing of
leaves could therefore be registered. The total amount of
work to produce dry Moringa leaves from 1 ton of fresh
Moringa was therefore 3.4 man days giving approximately
120 kg of Moringa leaves.
4. Discussion
The method used in the present experiment for the
production of MLM was a simple and cheap method of sun-
drying on plastic sheets and this method can easily be
adopted by small farmers. There are several other methods of
producing MLM such as freeze-drying (Richter et al., 2003),
drying in the shadow (Sarwatt et al., 2004 and Kakengi et al.,
2007) and combined air- and oor drying (Murro et al., 2003)
but the sun-drying method has been studied and proved to be
simple and adequate (Olsson and Wilgert, 2007).
It is important to realize that the chemical composition of
MLM can vary considerably mainly depending on the amount
of smaller branches and twigs included along with the leaves
in the leaf meal. This was shown by Fujihara et al. (2005),
who analyzed different fractions of Moringa (leaves, seed
cake, soft twigs, and bucks). The leaves and seed cake had a CP
content of approximately 250300 g kg
1
DM while leaves
with soft twigs had a CP content of 195 g kg
1
DM. The CP
content of soft twigs alone was yet somewhat lower but this
fraction can be used for animals with lower nutrient
requirements such as dry cows, who readily consume this
fraction. The drying method used in the present experiment
allows dried leaves to be easily removed from the coarser
fraction, giving a highly digestible product with a high
nutrient content. The MLM used here had an NDF content
of 161 g kg
1
DM. This value is similar to the NDF content
159 g kg
1
DM in the experiment of Richter et al. (2003)
where only leaves were used and much lower than the NDF
Table 3
Least square means of intake and apparent digestibility of dairy cows fed different concentrates.
Items Treatments SE Signicance
level
MLM concentrate Iso concentrate SBM concentrate
Feed intake (kg DM day
1
)
Elephant grass 9.45 9.22 10.38 0.44 ns
Concentrate 5.82 5.54 5.92 0.08 ns
Nutrient intake kg day
1
Dry matter 15.27 14.76 16.30 0.42 ns
Organic matter 13.62 13.41 14.58 0.37 ns
Crude protein 1.21
b
1.16
b
1.66
a
0.02 ***
Neutral detergent bre 7.07 6.92 7.52 0.30 ns
Acid detergent bre 5.70 5.73 6.20 0.05 ns
Lignin 1.15 1.13 1.22 0.04 ns
Starch 1.96
b
2.56
a
2.10
b
0.04 *
Metabolizable energy, MJ day
1
142
b
138
b
153
a
2.97 *
Apparent digestibility coefcient
Dry matter 0.74 0.74 0.73 0.01 ns
Organic matter 0.77 0.77 0.76 0.01 ns
Crude protein 0.67
ab
0.62
b
0.70
a
0.02 *
Neutral detergent bre 0.69 0.70 0.67 0.01 ns
Acid detergent bre 0.68 0.68 0.66 0.02 ns
abc
within a row means without common superscript differs. MLM: moringa leaf meal, SBM: soybean meal, SE: standard error, ns: not signicant.
Table 4
Least square means of milk yield and milk composition for cows fed different concentrates.
Items Treatments S.E. Signicance
level
MLM concentrate Iso concentrate SBM concentrate
Milk (kg day
1
) 12.3
b
12.1
b
13.2
a
0.20 *
Energy corrected milk (kg day
1
) 11.6
b
11.3
b
12.4
a
0.19 *
Milk fat (g kg
1
milk) 35.3 34.5 34.9 0.51 ns
Total solids (g kg
1
milk) 126.6 125.5 126.2 0.08 ns
Non fat solids (g kg
1
milk) 91.3 91.0 91.2 0.05 ns
Milk crude protein (g kg
1
milk) 34.7 34.3 34.6 0.23 ns
Casein (g kg
1
milk) 27.6 27.2 27.4 0.22 ns
MLM: moringa leaf meal, SBM: soybean meal, S.E.: standard error.
14 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
content 306 g kg
1
DM reported by Murro et al. (2003)
where twigs were included in the meal. Furthermore, the
simple sun-drying method used here does not seem to
decrease the CP concentration and the MLM used in the
present experiment was 292 g kg
1
DMwhich is in the range
of 250297 g kg
1
DM reported in other studies where the
leaf meal is produced almost entirely from the leaf fraction
(Richter et al., 2003; Kakengi et al., 2007). Moringa seems to
be a promising feed resource with a high nutrient content.
However, the in vitro studies of Fujihara et al. (2005) showed
that although the CP content in MLM was high with a high
rumen degradability, the proportion of protein potentially
degrade in the lower tract was lower compared to L.
leucocephala, another interesting protein source used in
tropical regions. It would therefore be valuable if in vivo
studies could be initiated in the future to study protein
availability on Moringa diets in more detail.
The CP content in the Elephant grass used in this
experiment, 34 g kg
1
DM, is below the range of 49 to
82 g kg
1
DM reported for tropical countries (Mendieta-
Araica et al., 2009; Sarwatt et al., 2004). However, in tropical
regions, animals are often fed roughages with very low
protein contents, similar to that of the elephant grass used in
the present experiment (Cndido et al., 2007 and Cavali et al.,
2010).
The CP content in the MLM concentrate and Iso concen-
trate diets were especially low; indeed, far below 143 g kg
1
DMas is recommended by the NRC (2001) for small breeds in
early to mid lactation producing 12.5 kg milk day
1
. Under
the production conditions for farmers who cannot afford to
buy concentrate supplements, these protein levels are,
however, common. Although diets with low protein content
have been reported to depress intake (M'hamed et al., 2001)
the DM intake in the present experiment was comparatively
high. In fact the lowCP content in the diets in this experiment
was to some extent compensated by high DM intakes.
Estimated intake by NRC (2001) for small breeds in early to
mid lactation with 12.5 kg ECM is 11.2 kg DM per day and
combined with the recommended CP content it renders a CP
intake per day of 1.6 kg for 12.5 kg ECM milk production. In
this experiment the CP allotment was 104% compared to the
NRC recommendation for the SBM diet, 76% for the MLM diet
and 73% for Iso diet. Although no signicant differences in
intake were observed in the present experiment, it is notable
that total intake was approximately 11.5 kg DM lower in the
treatments where animals were fed diets with a lower CP
content (i.e. the MLM concentrate and Iso concentrate
treatments).
Forages with less than 80 gCP kg
1
DM are dened by
Leng (1990) as low quality forages. Although the CP content
presented in Table 2 shows that the Elephant grass used in
this experiment was a low quality forage, the NDF content
was within the expected range of 521 to 784 g kg
1
DM
reported by Mendieta-Araica et al. (2009) and Sarwatt et al.
(2004) and together with a high bre digestibility this could
be the reason that intake of elephant grass was as expected.
The chemical composition of the MLM and Iso concen-
trates was similar with the exception of a higher starch
content in the Iso concentrate, which can be attributed to its
high proportion of sorghum. The difference between the
nutrient content of the SBM concentrate and the other
concentrate mixtures was, as expected, substantial, mainly
with regard to CP content but also to a certain extent with
regard to energy content.
The DM and organic matter digestibilities averaged 0.74
and 0.77, respectively, without signicant differences among
treatments, which is within the range of 0.70 to 0.75 for DM
digestibility reported by Murro et al. (2003) when MLM
replaces cottonseed cake as the protein source in concen-
trates for growing sheep and Nouala et al. (2006) when MLM
replaces 25% of commercial concentrates in an in vitro gas
production study. The lack of signicant differences in DM,
OM, NDF and ADF digestibility among treatments might be
due to the similar DM digestibility values of MLM and SBM,
which have been reported as 0.82 and 0.81, respectively
(Sarwatt et al., 2004; Loerch et al., 1983). Although there was
a difference in CP digestibility between the SBM and Iso
concentrate treatments, no difference in CP digestibility was
found between SBM and MLM concentrates, thus showing
that the CP in the Moringa diet was as digestible as in the
soybean diet.
The MLM and Iso diets in the present experiment had a
low CP content. This was largely due to the forage used. The
forage to concentrate ratio was also chosen to represent
proportions commonly used by farmers in Central America.
Therefore, the results of this experiment are directly relevant
for the farming situation of the region.
Milk yield was higher (Pb0.05) from cows fed the SBM
concentrate compared with the other concentrates (Table 4).
However, the difference was moderate: approximately 7%
lower when cows were fed the MLM concentrate (Table 4). It
is the local prices of SBM compared with milk that will
determine whether or not MLM is an economical alternative
to SBM in concentrates. The difference in milk yield can be
explained by the higher ME and CP intake when cows were
fed the SBM concentrate. There were no signicant differ-
ences in either milk yield or ECM between MLM and Iso
concentrates, indicating that the protein sources used in these
two concentrate mixtures were of similar quality with regard
to milk production. Milk composition was similar in all of the
treatments, which is consistent with previous studies
showing that low protein diets for dairy cows have little or
no effect on the fat and protein content in milk (Frank and
Swensson, 2002; Nielsen et al., 2003).
There is a common opinion among farmers that dairy
cows fed fresh Moringa will produce milk with a bad taste or
smell. Makkar and Becker (1997) attribute this bitter taste in
fresh Moringa to alkaloids, saponins and glucosinolates. The
results of the organoleptic analysis of milk in this experiment
showed no evidence of quality problems for any of the
treatments. The values for taste, smell and colour were all
typical, with no signicant differences among treatments. It is
possible that the earlier mentioned substances disappeared in
the drying and storage process used to produce the leaf meal.
Moringa has been reported to give a fresh matter yield of
71.4 Mg ha
1
year
1
corresponding to a dry matter yield of
13.5 Mg ha
1
year
1
at 45 days of cutting frequency (Reyes-
Snchez et al., 2006). Even though this was at planting
densities of 250,000750,000 plants per ha, which is higher
than the 100,000170,000 plants per ha used in the present
experiment, it shows that the production potential of
Moringa is substantial. The amount of work to produce
15 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
approximately 120 kg dried Moringa leaves was 34 man
days in the present experiment and the method is therefore
interesting in situations where labour is available but cash
money is scarce. It is therefore possible for small farmers in
marginal areas to cultivate Moringa and produce leaf meal
themselves to use as a supplement to their animals during the
dry season when there is a crucial shortage of protein in the
available feed. Still, more studies with MLM in other common
tropical diets are needed to obtain more detailed results with
regard to howfeeding Moringa affects dairy production under
conditions where protein quantity and quality are limiting
factors.
5. Conclusion
Moringa leaf meal is a potential source of protein to
supplement poor-quality forage such as Elephant grass. It can
successfully replace commercial concentrate constituents for
dairy cows as long as the substitution is isocaloric and
isoproteinic.
Acknowledgments
The funding for this research provided by the Swedish
International Development Agency (SIDA) is gratefully
acknowledged. Ing. Pablo Valdivia, the owner of Santa Ana
Farm, is also acknowledged for his help and support
throughout the experimental period.
References
Anhwange, B., Ajibola, V., Oniye, S., 2004. Amino acid composition of the
seeds of Moringa oleifera (Lam), Detarium microcarpum (Guill & Sperr)
and Bauhinia monandra (Linn.). Chem. Class J. 2004, 913.
AOAC, 1984. Ofcial Methods of Analysis, 14th edition. Association of ofcial
analytical chemists, Washington, US.
AOAC, 1990. Ofcial Methods of Analysis, 15th edition. Association of ofcial
analytical chemists, Gaithersburg, US.
Aregheore, E., 2002. Intake and digestibility of Moringa oleiferabatiki grass
mixture by growing goats. Small Ruminant Res. 46, 2328.
Becker, K., 1995. Studies on utilization of Moringa oleifera leaves as animal
feed. Inst. Anim. Prod. Tropics Subtropics 480, 15.
Broderick, G., Bradford, R., Spence, L., 1990. Expeller soybean meal and corn
by-products versus solvent soybean meal for lactating dairy cows fed
alfalfa silage as sole forage. J. Dairy Sci. 73, 453462.
Cndido, M., Neiva, J., Rodriguez, N., Ferreira, A., 2007. Fermentative patterns
and chemical composition of elephant grass silages with dehydrated
passionfruit byproduct. R. Bras. Zootec. 36, 14891494.
Castillo, A., Kebread, E., Beever, D., Barbi, J., Sutton, J., Kirby, H., France, J.,
2001. The effect of protein supplementation on nitrogen utilization in
lactating dairy cows fed grass silage diets. J. Anim. Sci. 79, 247253.
Castro Ramrez, A., 2002. Ganadera de leche: Enfoque empresarial.
Universidad Estatal a Distancia, Sn Jos, Costa Rica pp 287.
Cavali, J., Pereira, O., Valadares, S., Porto, M., Fernandes, F., Garcia, R., 2010.
Mixed sugarcane and elephant grass silages with or without bacterial
inoculant. R. Bras. Zootec. 39, 462470.
de Leeuw, P., Omore, A., Staal, S., Thorpe, W., 1999. Dairy production systems
in the tropics. In: Falvey, L., Chantalakkhana, C. (Eds.), Smallholder
Dairying in the Tropics. ILRI, Nairobi, pp. 1944.
Duckworth, J., Woodham, A., 1961. Leaf protein concentrates. I.Effect of
source of raw material and method of drying on protein value for chicks
and rats. J. Sci. Food Agric. 12, 515.
Fasuyi, A., Fajemilehin, S., Omojola, A., 2005. The egg quality characteristics of
layers fed varying dietary inclusions of Siam weed (Chromolaena
odorata) leaf meal (SWLM). Int. J. Poult. Sci. 4, 752757.
Frank, B., Swensson, C., 2002. Relationship between content of crude protein
in rations for dairy cows and milk yield, concentration of urea in milk and
ammonia emission. J. Dairy Sci. 85, 18291838.
Franziska, P., Baccini, P., 2005. Resources potentials and limitations of a
Nicaraguan agricultural region. Environ. Dev. Sust. 7, 337361.
Fujihara, T., Kakengi, A., Shem, M., Sarwatt, S., 2005. Can Moringa oleifera be
used as a protein supplement for ruminants? Asian-Aust. J. Anim. Sci.
18 (1), 42.
Ghosh, M., Atreja, P., Buragohain, R., Bandyopadhyay, S., 2007. Inuence of
short-term Leucaena leucocephala feeding on milk yield and its
composition, thyroid hormones, enzyme activity, and secretion of
mimosine and its metabolites in milk cattle. J. Agric. Sci. 145, 407414.
Gidamis, A., Panga, J., Sarwatt, S., Chove, B., Shayo, N., 2003. Nutrients and
antinutrients contents in raw and cooked leaves and mature pods of
Moringa oleifera Lam. Ecol. Food Nutr. 42, 113.
Hammond, A., 1995. Leucaena toxicosis and its control in ruminants. J. Anim.
Sci. 73, 14781492.
Hazard, S., 1990. Sabe usted como alimentar sus vacas lecheras. Investigacin
y Progreso Agropecuario Carillanca. 9, 38-41 In: INTA (ed.) 2000. Manual
Tcnico de Extensin Pecuaria. Managua, Nicaragua.
Kakengi, A., Shem, M., Mtengeti, E., Otsyina, R., 2001. Leucaena leucocephala
leaf meal as supplement to diet of grazing dairy cattle in semiarid
Western Tanzania. Agrofor. Syst. 52, 7382.
Kakengi, A., Kaijage, J., Sarwatt, S., Mutayoba, S., Shem, M., Fujihara, T., 2007.
Effect of Moringa oleifera leaf meal as a substitute for sunower seed
meal on performance of laying hens in Tanzania. Livest. Res. Rural Dev.
19, 120.
Leng, R., 1990. Factors affecting the utilization of poor quality forages by
ruminants particularly under tropical conditions. Nutr. Res. Rev. 44,
277303.
Lindgren, E., 1979. Vallfodrets nringsvrde bestmt in vivo och med olika
laboratoriemetoder. Report 45 The Department of Animal Nutrition and
Management. The Swedish University of Agricultural Sciences, Uppsala,
Sweden.
Loerch, S., Berger, L., Plegge, D., Fahey, G., 1983. Digestibility and rumen
escape of soybean meal, blood meal, meat and bone meal and
dehydrated alfalfa nitrogen. J. Anim. Sci. 57, 10371047.
Ly, J., Samkol, P., Preston, T., 2001. Nutritional evaluation of tropical leaves for
pigs: pepsin/pancreatin digestibility of thirteen plan species. Livest. Res.
Rural Dev. 13, 5.
M'hamed, D., Faverdin, P., Verit, R., 2001. Effects of the level and source of
dietary protein on intake and milk yield in dairy cows. Anim. Res. 50,
205211.
MAGFOR, 2008. Informe anual de estadstico de produccion agropecuaria por
ciclo agrcola. ((Annual statistic report of agricultural production by
agricultural cycle)). Ministerio agropecuario y forestal, Managua,
Nicaragua.
Makkar, H., Becker, K., 1997. Nutrients and antiquality factors in different
morphological parts of the Moringa oleifera tree. J. Agric. Sci. 128,
311332.
Mapiye, C., Chimonyo, M., Dzama, K., Strydom, P., Muchenje, V., Marufu,
2009. Nutritional status, growth performance and carcass characteristics
of Nguni steers supplemented with Acacia karroo leaf-meal. Livest. Sci.
126, 206214.
McDonald, P., Edwards, R., Greenhalgh, J., 1988. Animal Nutrition, fourth ed.
Longman group, United Kingdom.
McDonald, P., Edwards, R., Greenhalgh, J., Morgan, C., 2002. Animal Nutrition,
sixth ed. Longman group, United Kingdom.
Mendieta-Araica, B., Sprndly, E., Reyes-Snchez, N., Norell, L., Sprndly, R.,
2009. Silage quality when Moringa oleifera is ensiled in mixtures with
Elephant grass, sugar cane and molasses. Grass Forage Sci. 64, 364373.
Murro, J., Muhikambele, V., Sarwatt, S., 2003. Moringa oleifera leaf meal can
replace cottonseed cake in the concentrate mix fed with Rhodes grass
(Chloris gayana) hay for growing sheep. Livest. Res. Rural Dev. 15, 11.
Nielsen, N., Kristensen, T., Norgaard, P., Hansen, H., 2003. The effect of low
protein supplementation to dairy cows grazing clover grass during half
of the day. Liv. Prod. Sci. 81, 293306.
Nouala, F., Akinbamijo, O., Adewumi, A., Hoffman, E., Muetzel, S., Becker, K.,
2006. The inuence of Moringa oleifera leaves as substitute to
conventional concentrate on the in vitro gas production and digestibility
of groundnut hay. Livest. Res. Rural Dev. 18, 9.
NRC, 2001. Nutrient Requirements of Dairy Cows, 7th rev. ed. Nat. Acad. Sci,
Washington, DC.
Olafadehan, O., Adewumi, M., 2008. Milk production and economic impact of
strategic supplementation of prepartum Bunaji cows in the peri-urban
areas of derived savannah of southwestern Nigeria. Livest. Res. Rural
Dev. 20, 3.
Olafadehan, O., Adewumi, M., 2009. Productive and reproductive perfor-
mance of strategically supplemented free grazing prepartum Bunaji
cows in the agropastoral farming system. Trop. Anim. Health Prod. 41,
12751281.
Olafadehan, O., Adewumi, M., 2010. Milk yield and composition of prepartum
Bunaji cows supplemented with agroindustrial by-products in small-
holder dairy production systems. Trop. Subtrop. Agroecosyst. 12,
557564.
16 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
Olsson, L., Wilgert, E., 2007. Moringa oleiferaan evaluation of drying
methods to produce dry season feed for cattle. Minor eld studies No 390
SLU External relations. Swedish University of Agricultural Sciences,
Uppsala, Sweden.
Paterson, R., Karanja, G., Roothaert, R., Nyaata, O., Kariuki, I., 1998. A reviewof
tree fodder production and utilization within smallholder agroforestry
systems in Kenya. Agrofor. Syst. 41, 181199.
Patra, A., Sharma, K., Dutta, N., Pattanaik, A., 2003. Response of gravid does to
partial replacement of dietary protein by a leaf meal mixture of Leucaena
leucocephala, Morus alba and Azadirachta indica. Anim. Feed Sci. Technol.
109, 171182.
Patterson, H., Lucas, H., 1962. Change-over design. Tech. Bull. 147 North
Carolina. US.
Pereira, F., 1988. Alimentos. Manual de anlisis fsico qumico. Universidad
Autnoma de Yucatn. Direccin general de diffusin y communicacin.
Mrida Yucatn, Mexico, pp. 190.
Reyes-Snchez, N., Sprndly, E., Ledin, I., 2006. Effect of feeding different
levels of foliage of Moringa oleifera to creole dairy cows on intake,
digestibility, milk production and composition. Livest. Sci. 101, 2431.
Richter, N., Siddhuraju, P., Becker, K., 2003. Evaluation of nutritional quality
of Moringa (Moringa oleifera Lam.) leaves as an alternative protein
source for Nile tilapia. Aquaculture 217, 599611.
Romero, F., Gonzlez, J., 2004. Effects of dry season feeding of fresh and
ensiled Cratylia argentea on milk production and composition. In:
Holmann, F., Lascano, C. (Eds.), Feeding systems with forage legume to
intensify dairy production in Latin America and the Caribbeann.
Tropileche consortium.
Snchez-Machado, D., Nez-Gastlum, J., Reyes-Moreno, C., Ramrez-Wong,
B., Lpez-Cervantes, J., 2010. Nutritional quality of edible parts of
Moringa oleifera. Food analytical methods 3 (3), 175180.
Sarwatt, S., Milang'ha, M., Lekule, F., Madalla, N., 2004. Moringa oleifera and
cottonseed cake as supplements for smallholder dairy cows fed Napier
grass. Livest. Res. Rural Dev. 16, 6.
SAS, 2004. Statistical Analysis System Institute Inc. User Guide, version 9.1.2.
U.S.
Shem, M., Macibula, B., Sarwatt, S., Fujihara, T., 2003. Gliricidia sepium as an
alternative protein supplement to cottonseed cake for smallholder dairy
cows fed on Napier grass in Tanzania. Agrofor. Syst. 58, 6572.
Undersander, D., Mertens, D., Theix, N., 1993. Forage Analyses Procedures.
National forage testing association, Omaha, U.S.
USDA, 2009. Oilseeds: World Markets and Trade. United State Department of
Agriculture. Circular series FOP 1-09, Jan. 2009.
Van Soest, P., Robertson, Lewis, B., 1991. Methods for dietary ber, neutral-
detergent ber and non-starch polysaccharides in relation to animal
nutrition. J. Dairy Sci. 4, 35833597.
Vlez, M., 1997. Produccin de ganado lechero en el Trpico. Escuela Agrcola
Panamericana, El Zamorano. Tegucigalpa, Honduras, pp. 189.
Witting de Penna, E., 1981. Evaluacin sensorial, una metdica que mide
calidad. II. Evaluacin mediante el test de valoracin con escala de
Karlsruhe. Alimentary 6, 2531.
17 B. Mendieta-Araica et al. / Livestock Science 137 (2011) 1017
II
Silage quality when Moringa oleifera is ensiled
in mixtures with Elephant grass, sugar cane
and molasses
B. Mendieta-Araica*, E. Spo rndly*, N. Reyes-Sanchez, L. Norell and R. Spo rndly*
*Department of Animal Nutrition and Management, Swedish University of Agricultural Sciences, Uppsala,
Sweden, Faculty of Animal Science, Universidad Nacional Agraria, Managua, Nicaragua, and Unit of Applied
Statistics and Mathematics, Department of Economics, Swedish University of Agricultural Sciences, Uppsala,
Sweden
Abstract
Fourteen different silages were prepared using mixtures
of Moringa (Moringa oleifera), Elephant grass (Pennise-
tum purpureum cv Taiwan) or sugar cane (Saccharum
ofcinarum). Molasses from sugar cane was used in the
amounts of either 10 or 50 g kg
)1
fresh matter (FM) in
treatments without sugar cane. A completely random-
ized design with three replicates of each treatment was
used. The silages were prepared in 1800 mL micro silos
and opened after 120 d. The presence of Moringa and
Elephant grass in the silage changed the pH by )08 and
+07, respectively (P < 0001), indicating a favourable
effect of Moringa on silage pH. Overall differences were
found among treatments for dry matter content, crude
protein and acetic acid concentrations, weight loss, CO
2
production and silage pH after spoilage (P < 0001).
Weight loss was proportionately 0034 and 0014 in
silages with and without sugar cane respectively
(P < 0001). Overall, differences (P < 005) were also
found for neutral-detergent bre and lactic acid con-
centrations, lactic acid bacteria counts, clostridial counts
and time to spoilage of the silages. Treatments contain-
ing Moringa had higher lactic acid concentrations
(+16 g kg
)1
DM; P < 001) compared to treatments
without but the presence of Moringa decreased time
to spoilage by 67 h (P < 005). No differences were
found in propionic acid concentration or fungal growth
of the silages. It is concluded that Moringa can be used
as a component of high quality silages which also
contain high concentrations of crude protein.
Keywords: Moringa oleifera, Elephant grass, sugar cane,
tropical feedstuffs, ensiling, spoilage
Introduction
In tropical countries where livestock production is
mainly based on grass-dominated pastures, herbage
mass during the dry season is generally not sufcient
to satisfy the nutritional requirements of livestock. To
mitigate this, silage is one of the alternative feeds as it
is relatively simple to produce and utilizes the surplus
in herbage production from the rainy season. Elephant
grass (Pennisetum spp) is commonly used for silage but
authors report that the herbage has a low content of
dry matter (DM) and low concentrations of water
soluble carbohydrates (WSC) and crude protein (CP)
(Andrade and Melotti, 2004; Mtengeti et al., 2006;
Zanine et al., 2006). The composition of the herbage
makes it difcult to produce high quality silage
without the use of additives. Adding molasses, or other
sources of WSC, is widely used to promote a low pH
and high proportions of lactate in such silages (Uman a
et al., 1991; Andrade and Melotti, 2004; Mtengeti
et al., 2006). Silage made from sugar cane (Saccharum
ofcinarum), either pure or mixed with grasses, is also
used in tropical areas. Undesired ethanol production,
however, leads to substantial DM losses and low
quality silage when pure sugar cane is ensiled (Pedroso
et al., 2005).
The low quality of tropical silages is often caused by
the low nutritive value of the herbage ensiled. McDo-
well (1972) studied 312 tropical and 760 temperate
species and found that in more than half of the tropical
grasses the concentration of total digestible nutrients
(TDN) was up to 15 units less than that of temperate
grasses. The need for higher quality silage in tropical
areas calls for new solutions using unconventional
Correspondence to: R. Spo rndly, Department of Animal
Nutrition and Management, Swedish University of Agri-
cultural Sciences, P.O. Box 7024, 75007 Uppsala, Sweden.
E-mail: rolf.sporndly@huv.slu.se
Received 5 February 2009; revised 18 June 2009
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373 doi: 10.1111/j.1365-2494.2009.00701.x
364
forage species and, according to Ca rdenas et al. (2003),
the use of tree foliage as a feed for livestock has
increased. Tree foliage can be used to obtain silage with
higher CP concentrations and offers the possibility to
replace conventional concentrates (Ca rdenas et al.,
2003).
While there are many agro-forestry species of inter-
est, one of the most interesting trees is Moringa oleifera,
commonly referred to as Moringa. It is one of the most
widely utilized species (Makkar and Becker, 1996,
1997). Moringa is a fast-growing tree which can reach
12 m in height at maturity and yield up to 88 t ha
)1
fresh matter (FM) annually when planted very densely
for use as a forage. The CP concentration in leaves is
about 200250 g kg
)1
DM with a negligible amount of
tannins in all fractions of the Moringa plant and
high levels of sulphur-containing amino acids (Reyes-
Sa nchez et al., 2006).
The principal aim of this experiment was to evaluate
the effect on fermentation characteristics when Moringa
is introduced for ensiling in various mixtures with at
least one of the components of Elephant grass, sugar
cane or sugar-cane molasses. Particular attention was
paid to losses and the aerobic stability of silages. The
hypothesis tested was that the introduction of Moringa
oleifera, as a silage component to improve the nutri-
tional value of silages used for supplementary feeding in
the tropics, will produce silages with fermentation
characteristics and aerobic stability that are comparable
to traditional silages based on pure sugar cane or
Elephant grass with molasses as an additive.
Materials and methods
This experiment was carried out at the National
University of Agriculture (UNA) in Managua, Nicaragua
between March and June 2007. In the laboratory, the
average relative humidity was 551% and the average
temperature was 315C, with the highest temperatures
occurring towards the end of April and a large variation
in both humidity and temperature.
Micro-silo preparation
Non-irrigated, unfertilized Moringa (Moringa oleifera)
foliage (leaves and branches of <5 mm diameter),
Elephant grass (Pennisetum purpureum cv Taiwan) and
sugar cane (Saccharum ofcinarum) were harvested from
the experimental elds of UNA. Sugar-cane molasses
(molasses) was purchased from an agricultural feed
supplier. The chemical composition of the materials
used is presented in Table 1.
All forages were hand-cut with a machete. Moringa
and Elephant grass were cut 45 d after pruning. Sugar
cane was cut 9 months after the previous cut. All leaves
and roots were removed from the stems and only the
stems were ensiled. Due to the low nutrient content of
the sugar-cane leaves, it is common practice to remove
the leaves along with the roots before ensiling. After
cutting, forages were chopped into pieces of approxi-
mately 2 cm in length using a mechanical chopper.
Once the forages were chopped, fourteen treatments
were prepared from different proportions of Moringa,
Elephant grass and sugar cane for ensiling (see Table 2).
In treatments without sugar cane, molasses without any
Table 1 Dry matter (DM) content and concentrations of
crude protein (CP), water-soluble carbohydrates (WSC),
neutral-detergent bre (NDF), acid-detergent bre (ADF) and
lignin of ensiled feeds.
Feeds
Moringa
Elephant
grass
Sugar
cane
Sugar-cane
molasses
DM content
(g kg
)1
)
193 197 223 721
Concentration
(g kg
)1
DM) of:
CP 268 49 27 22
Ash 15 18 15 28
WSC <50 <50 81 472
NDF 521 737 669 ND
ADF 361 374 380 ND
Lignin 119 102 63 ND
ND, not determined.
Table 2 Proportions of Moringa (M), Elephant grass (E), sugar
cane (Sc) and molasses of sugar cane (Sm) in the treatments.
Feeds
Treatment
designation
Moringa
(g kg
)1
)
Elephant
grass
(g kg
)1
)
Sugar
cane
(g kg
)1
)
Molasses
(g kg
)1
)
M99E0Sm1 990 10
M95E0Sm5 950 50
M67E0Sc33 667 333
M66E33Sm1 660 330 0 10
M63E32Sm5 633 317 0 50
M33E33Sc33 333 333 333
M33E66Sm1 330 660 10
M32E63Sm5 317 633 50
M33E0Sc67 333 667
M0E99Sm1 990 10
M0E95Sm5 950 50
M0E67Sc33 667 333
M0E33Sc67 333 667
M0E0Sc100 1000
Silages made from tropical grasses and Moringa
365
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
water dilution was added at rates of 10 or 50 g kg
)1
FM.
Mixtures for each treatment were prepared and from
these batches fresh material was taken to ll on average
1564 g FM in glass jars with a nominal volume of
1800 mL. The fresh material was pressed into the jar to
remove as much air as possible. The weight of the fresh
material in each jar was calculated from the difference
between empty and lled jars. The glass jars, subse-
quently referred to as micro-silos, were tted with
water-locks on their lids to let fermentation gases
escape. Each of the fourteen treatments had three
replicates, giving a total of forty-two micro-silos. The
temperature and relative humidity of the store room
was monitored three times daily. The water-locks were
relled when needed. Micro-silo weight was recorded
every 2 d at 08:00 h.
All micro-silos were opened and analysed 120 d after
being closed. The contents from each micro-silo were
transferred into a plastic bag, thoroughly mixed and
then samples taken for analysis. The DM content of
each sample was determined by oven-drying at 105C
for 12 h. Concentrations of ash, CP (as 625 N
concentration) and WSC were determined as described
by AOAC (2000). A Thermo Scientic Orion 2-Star
Benchtop pH meter (Thermo FisherScientic Inc.,
Waltham, MA, USA) was used to determine pH. The
concentrations of neutral-detergent bre (NDF) and
acid-detergent bre (ADF) were determined according
to Van Soest et al. (1991). Concentrations of lactic acid,
propionic acid and acetic acid were determined by High
Performance Liquid Chromatography (HP 1100, Agilent
Technologies, Santa Clara, CA, USA) with 40 mL silage
uid. Before being injected into the chromatograph, the
samples were centrifuged for 10 min at a temperature
of approximately 4C to prevent loss of volatiles.
Clostridium perfringens, lactic acid bacteria (LAB), aerobic
enterobacteria and fungal growth were determined
according to APHA (2001).
Aerobic stability phase
From each micro-silo, a 250-g sample was taken to
evaluate aerobic stability using the method of Ashbell
et al. (1990) to measure CO
2
production as an estimate
of microbial activity. These samples were aerobically
stored and protected by nets against insects. Both the
sample and ambient temperature were measured
during the deterioration process every 2 h.
Once a sample temperature of 5C above room
temperature was recorded three times in a row, the
sample was considered spoiled and sent, together with
its potassium hydroxide solution, for chemical analysis.
The concentration of CO
2
was measured according to
Ashbell et al. (1990). From 10 g, blended for 5 min in a
laboratory blender with 90 mL distilled water, ltrate
pH was determined using a Thermo Scientic Orion
2-Star Benchtop pH meter.
Experimental design and statistical analysis
A completely randomized experimental design was
used, with fourteen treatments and three replicates
for each treatment. The treatments are presented in
Table 2 and, as can be seen in the table, the design was
not symmetric with regard to proportions of compo-
nents. Small amounts of sugar-cane molasses (10 or
50 g kg
)1
) were included in silages with Moringa and
Elephant grass but not in mixtures with sugar cane. The
reason was that ensiling of sugar cane using sugar-cane
molasses as an additive rarely occurs in practice. Each of
these two components generally contains readily
fermentable WSC, which are favourable for the ensiling
process, and it is regarded uneconomical and unneces-
sary to add molasses when the mixture already contains
sugar cane. A symmetric design would have included a
number of treatments with both sugar cane and sugar-
cane molasses.
A preliminary analysis including the DM content as a
covariate was undertaken. The variable did not improve
the analysis and was omitted. Thus, the variables were
analysed by the one-way model:
y
ij
l s
i
e
ij
; i 1; . . . ; 14; j 1; 2; 3
where l = overall mean, s
i
= treatment effect and
e
ij
N(0,r
2
) is a random error following the normal
distribution. In addition to the overall F-test of treat-
ment effects, pair-wise differences were studied. A
difference was considered as signicant if P 005,
where the P-values were adjusted for multiple compar-
isons according to Tukeys method.
The averages of the groups of treatments including a
feed component were compared with averages of the
treatments without the same component to summarize
the effect of including that component in the silage
mixture. It should be remembered that the results of
these comparisons are somewhat uncertain as the out-
come is also inuenced by the proportions and combi-
nations of the other components used in the two groups.
To enable an analysis of effects of the proportion of
silage components, mixture experiment models were
investigated, see e.g. Atkinson and Donev (1992). The
rst-order model was
y
ij
b
1
x
i1
b
2
x
i2
b
3
x
i3
b
4
x
i4
e
ij
;
i 1; . . . ; 14; j 1; 2; 3
where x
i1
,,x
i4
denote the proportions of Moringa,
Elephant grass, sugar cane and molasses respectively. In
addition to the usual restrictions x
i1
+ + x
i4
= 1 and
x
ik
0, the special restrictions x
i3
+ x
i4
001, x
i4
005
and x
i3
x
i4
= 0 were added, cf. rst paragraph in this
366
B. Mendieta-Araica et al.
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
section. The second-order model for the mixture
experiment was
y
ij
b
1
x
i1
b
4
x
i4
b
12
x
i1
x
i2
b
13
x
i1
x
i3
b
14
x
i1
x
i4
b
23
x
i2
x
i3
b
24
x
i2
x
i4
e
ij
The product x
i3
x
i4
is excluded from the model since
x
i3
x
i4
= 0 implying that b
34
is not possible to estimate.
Goodness-of-t tests were performed by using the
difference of the sums of squares of the one-way model
and of the mixture experiment model as the numerator
in an F-test. In cases where the second-order model was
applied, the product effects were successively tested and
removed until only signicant products remained. In
parallel with signicance testing of mixture experiment
models, descriptive correlation coefcients of the pre-
dicted values and the treatment means were calculated.
Variables not showing sufcient goodness-of-t with
the mixture experiment models were analysed by the
one-way model only.
The procedure of GLM in SAS (2004), including the
options MEANS, PDIFF and ESTIMATE, was used for
the numerical calculations.
Results
Chemical composition of silages
The DM content and concentrations of CP and NDF of
each silage treatment 120 d after ensiling are presented
in Table 3. In general, there was a narrow range in DM
content among treatments: 210269 g kg
)1
. Elephant
grass treatments had the highest DM contents regard-
less of their content of molasses.
Crude protein concentrations increased markedly
with increasing proportion of Moringa, with the highest
CP concentration being found at the highest proportion
of Moringa (treatment M99E0Sm1, 150 g kg
)1
DM)
and the lowest CP concentration with pure sugar cane
(treatment M0E0Sc100, 25 g kg
)1
DM). Furthermore,
the ANOVA ANOVA model estimated that treatment groups with
Moringa contained signicantly higher (+56 g kg
)1
DM; P < 0001) concentrations of CP compared to
treatment groups without Moringa, while the presence
of Elephant grass or sugar cane gave signicantly lower
CP concentrations, )35 and )30 g kg
)1
DM (P < 0001)
respectively. Treatment M66E33Sm1 had an unexpect-
edly low CP concentration (Table 3).
The concentration of NDF in silage decreased when
the proportion of Moringa increased. An inexplicably
low NDF concentration was recorded for treatment
M95E0Sm5, 397 g kg
)1
DM. It was not possible to
re-analyse the sample for NDF concentration to deter-
mine whether this was an error. All silage treatments
had WSC concentrations lower than the detection limit
of 50 g kg
)1
DM with the only exception being the
treatment composed entirely of sugar cane (treatment
M0E0Sc100) which had a WSC concentration of
71 g kg
)1
DM.
Fermentation prole
Silage pH, fermentation products and weight loss of
silage from the ANOVA ANOVA model are presented in Table 4.
The effect of treatment on silage pH (P < 001), lactic
and acetic acid concentrations (P < 001 and P < 0001
respectively), and weight loss (P < 0001), were signif-
icant in the model. DM content had no signicant effect
either on silage pH, fermentation products or weight
loss in the model and was, therefore, as mentioned in
the section on Experimental deign and statistical anal-
yses, excluded from the model.
When tested in the ANOVA ANOVA model, the presence of
Moringa in treatments decreased pH values by 08 (P <
00001). Correspondingly, the presence of Elephant
grass increased pH values by 07 (P < 0001). No such
effect was seen with the presence of sugar cane.
The presence of both Moringa and Elephant grass, as
well as the proportion of molasses, affected the lactic
acid concentration by 16 g kg
)1
DM (P < 0001),
)21 g kg
)1
DM (P < 0001) and )12 g kg
)1
DM
(P < 005) respectively. The presence of sugar cane
decreased acetic acid concentration (P < 005). When
sugar cane was the only silage component, the acetic
acid concentration was below the detection limit.
According to the ANOVA ANOVA analysis, there were signif-
icant differences (P < 0001) among treatments with
Table 3 Dry matter (DM) content and concentrations of
crude protein (CP) and neutral-detergent bre (NDF) of silage
treatments after ensiling for 120 days.
Treatments
DM
content
(g kg
)1
)
CP
concentration
(g kg
)1
DM)
NDF
concentration
(g kg
)1
DM)
M99E0Sm1 212 150 583
M95E0Sm5 217 144 397
M67E0Sc33 216 101 517
M66E33Sm1 240 67 518
M63E32Sm5 233 107 628
M33E33Sc33 235 71 622
M33E66Sm1 223 67 691
M32E63Sm5 253 68 562
M33E0Sc67 210 64 614
M0E99Sm1 264 41 689
M0E95Sm5 269 42 710
M0E67Sc33 258 45 776
M0E33Sc67 247 37 651
M0E0Sc100 237 25 740
Silages made from tropical grasses and Moringa
367
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
regard to weight loss of silage. The presence of sugar
cane caused a signicantly (P < 0001) greater propor-
tional weight loss from 0014 to 0034 (see Table 4).
Microbiological composition of silages
The microbiological composition of the silage treat-
ments is shown in Table 5. The highest proportion of
Elephant grass resulted in the lowest number of
clostridia (20 log cfu g
)1
). A tendency for higher
Clostridia numbers was also observed in treatments
with sugar cane rather than molasses. Intermediate
LAB numbers were found in treatments with Moringa
while the highest number (65 log cfu g
)1
, P < 001)
was observed with a combination of Elephant grass and
sugar cane (treatment M0E67Sc33) and the lowest with
a combination of Elephant grass and molasses (treat-
ment M0E95Sm5). There were no signicant differ-
ences in fungal growth among treatments. All silages
had Enterobacteria numbers below the detection limit
of log 05 MPN g
)1
.
Aerobic stability of silages
Results from the aerobic stability phase are presented in
Table 6. The production of CO
2
(P < 0001), time to
spoilage (P < 001) and pH after spoilage (P < 0001)
were all signicantly affected by treatment in the
model.
When the estimates were tested in the ANOVA ANOVA model,
the presence of Moringa decreased time to spoilage by
67 h (P < 005). The presence of Moringa caused a
signicant (P < 0001) increase in CO
2
production
while the presence of Elephant grass caused a decrease
(P < 0001). Increasing the proportion of molasses from
001 to 005 decreased time to spoilage by 103 h
(P < 005); however, treatment M66E33Sm1 had an
unexpectedly long time to spoilage and treatment
Table 4 The concentrations of lactic,
acetic and propionic acids, pH and weight
losses (WL) in silage treatments after
120 days of ensiling. Means from ANOVA ANOVA
model with standard error of mean and
level of signicance of treatments are also
given.
Treatments
Lactic acid
(g kg
)1
DM)
Acetic acid
(g kg
)1
DM)
Propionic acid
(g kg
)1
DM) pH
WL
(g kg
)1
)
M99E0Sm1 93
abc
31
ab
03 370
b
9
g
M95E0Sm5 106
a
26
ab
02 353
b
13
efg
M67E0Sc33 98
ab
17
bcd
70 349
b
24
cd
M66E33Sm1 79
abc
26
ab
28 406
ab
14
efg
M63E32Sm5 93
abc
24
ab
45 349
b
11
fg
M33E33Sc33 90
abc
3
d
51 371
b
24
c
M33E66Sm1 73
abc
17
bcd
18 509
ab
14
efg
M32E63Sm5 75
abc
15
bcd
55 399
ab
14
efg
M33E0Sc67 99
ab
2
d
45 354
b
38
b
M0E99Sm1 55
c
36
a
60 470
b
20
cde
M0E95Sm5 74
abc
21
abc
65 420
ab
19
def
M0E67Sc33 64
bc
4
cd
32 584
a
28
c
M0E33Sc67 85
abc
2
d
27 422
ab
41
ab
M0E0Sc100 92
abc
>1 27 356
b
47
a
s.e. of mean 57 31 16 008 017
P-value <0001 <0001 NS <001 <0001
Columns without common superscripts differ signicantly (P < 005); NS, not
signicant.
Table 5 Concentrations of clostridia, lactic acid bacteria
(LAB) and fungi of each silage treatment after 120 days of
ensiling with standard error of mean and levels of signicance of
treatments.
Treatments
Clostridia
(log cfu g
)1
)
LAB
(log cfu g
)1
)
Fungi
(log MPN g
)1
)
M99E0Sm1 38
ab
36
abc
13
M95E0Sm5 36
ab
25
abc
16
M67E0Sc33 32
ab
29
abc
11
M66E33Sm1 31
ab
24
abc
05
M63E32Sm5 35
ab
32
abc
08
M33E33Sc33 35
ab
35
abc
05
M33E66Sm1 34
ab
29
abc
05
M32E63Sm5 32
ab
30
abc
05
M33E0Sc67 33
ab
27
abc
05
M0E99Sm1 20
b
19
bc
05
M0E95Sm5 25
ab
08
c
05
M0E67Sc33 42
ab
65
a
13
M0E33Sc67 44
a
49
ab
05
M0E0Sc100 35
ab
19
bc
05
s.e. of mean 042 077 031
P-value <005 <001 NS
Columns without common superscripts differ signicantly
(P < 005); NS, not signicant.
368
B. Mendieta-Araica et al.
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
M63E32Sm5 had an unexpectedly short time. The
presence of Moringa caused a signicant (P < 0001)
increase in pH after spoilage; when the estimates were
tested in the ANOVA ANOVA model, the presence of Moringa
increased pH after spoilage by 155 (P < 0001).
Mixture experiment models
With the exception of fungi, all variables in Tables 46
were analysed in mixture experiment models and the
variables that showed goodness-of-t for rst or second
order models are presented in Table 7 along with
P-values and the descriptive correlation coefcients
with sample means. The results of the model estima-
tions must be considered while keeping in mind the
restriction of not having the components sugar cane
and molasses in the same mixture.
A mixture experiment model was considered relevant
if the P-value of the goodness-of-t test was greater
than 005 and the descriptive correlation coefcient
versus the sample means was at least 090. A second-
order model was used only if the rst- order model was
insufcient. The variables that showed a relevant t to
mixture experiment models are given in Table 7.
The predicted values based on the rst-order model
are given in Table 8. Under the restrictions for the
x values, described in the section on Experimental
design and statistical analyses, the estimated rst-order
model for lactic acid concentration has its maximum
at x
1
= 095, x
4
= 005, corresponding to treatment
M95E0Sm5, which also had the highest sample mean
(see Table 4). The minimum value according to the
rst-order model is attained at x
2
= 099, x
3
= 001, a
composition not included in the experiment. Among
the studied treatments (see Table 4), the lowest mean
lactic acid concentration of 55 g kg
)1
DM was obtained
for treatment M0E99Sm1 with x
2
= 099, x
4
= 001,
Table 6 Carbon dioxide production during the aerobic
stability phase, time to spoilage (TTS) and pH after spoilage of
the treatments with standard error of mean and level of
signicance.
Treatments
CO
2
production
(g kg
)1
DM)
TTS
(h) pH
M99E0Sm1 42
ab
81
abc
625
bcd
M95E0Sm5 43
a
115
abc
750
ab
M67E0Sc33 39
abc
106
abc
456
de
M66E33Sm1 34
bcd
303
abc
932
a
M63E32Sm5 39
abc
39
bc
551
cde
M33E33Sc33 34
bcd
194
abc
777
ab
M33E66Sm1 34
bcd
180
abc
773
ab
M32E63Sm5 31
cde
151
abc
675
bc
M33E0Sc67 31
cde
237
abc
413
e
M0E99Sm1 25
e
354
a
518
cde
M0E95Sm5 26
de
246
abc
462
de
M0E67Sc33 30
cde
33
c
500
cde
M0E33Sc67 28
de
191
abc
599
cde
M0E0Sc100 31
cde
314
ab
454
de
s.e. of mean 17 530 038
P-value <0001 <001 <0001
Columns without common superscripts differ signicantly
(P < 005).
Table 7 Variables [concentrations of clostridia, Lactic acid
bacteria (LAB), lactic acid concentration, weight loss (WL)
and CO
2
production] with goodness-of-t in rst or second
order mixture experiment models; p
1,
p
2
and r
1,
r
2
are the
P-values of goodness-of-t tests and descriptive correlation
coefcients of rst and second order models, respectively,
vs. sample means.
Variable p
1
r
1
p
2
r
2
Model order
chosen
Clostridia 0081 054 084 096 2nd
LAB 0005 038 028 092 2nd
Lactic acid 0805 095 071 097 1st
WL 0308 099 054 100 1st
CO
2
production 0227 095 057 098 1st
Table 8 Variables modelled by the rst order mixture
experiment model [ concentration of lactic acid, weight loss
(WL) and CO
2
production]. Estimated coefcients with
standard errors (s.e. of coefcient) for predicting equation y and
values of components x
1
(Moringa)
,
x
2
(Elephant grass)
,
x
3
(sugar cane) and x
4
(molasses) giving proportions of these
components at maximal (prop. at max.) and minimal (prop. at
min.) predicted values. Corresponding sample mean y is given
provided the composition exists in Table 2.
Variable x
1
x
2
x
3
x
4
y y
Lactic acid
Coefcient 938 550 985 3099
s.e. of coefcient 40 43 47 915
Prop. at max. 095 0 0 005 1046 1055
Prop. at min. 0 099 001 0 554 Non-
existent
WL
Coefcient 103 188 491 072
SE of coefcient 012 013 013 276
Prop.at max. 0 0 1 0 491 470
Prop. at min. 095 0 0 005 101 089
CO
2
production
Coefcient 417 258 302 531
SE of coefcient 12 13 13 276
Prop. at max. 095 0 0 005 423 431
Prop. at min. 0 099 001 0 258 Non-
existent
Silages made from tropical grasses and Moringa
369
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
which in the interpretation of the rst-order model is
very close to x
2
= 099, x
3
= 001.
The nal second-order mixture models after removal
of non-signicant product effects are presented in
Table 9. The maximum value of the variable clostrid-
ium, according to the second-order model, is obtained
at x
2
= 042, x
3
= 058 with a predicted value of 449 log
cfu g
)1
. The two treatments with the highest sample
means of 441 and 422 log cfu g
)1
(see Table 5)
correspond to x
2
= 1 3, x
3
= 2 3 and x
2
= 2 3,
x
3
= 1 3. These treatments with predicted values of
444 and 407 log cfu g
)1
are, in a mathematical sense,
close neighbours to x
2
= 042, x
3
= 058 and the max-
imum value of 449 log cfu g
)1
is a consequence of the
curvature due to the product effect of x
2
x
3
. For LAB, the
second-order model attains its maximum value of 598
log cfu g
)1
at x
2
= 050, x
3
= 050, a treatment not
included in the experiment. The closest treatment
neighbours, i.e. those with x
2
= 2 3, x
3
= 1 3 and
x
2
= 1 3, x
3
= 2 3, yielded the sample means y 6:50
log cfu g
)1
and y 4:92 log cfu g
)1
respectively. This is
contradictory. However, comparisons of the difference
between the second-order model predicted values, i.e.
^y 5:55 log cfu g
)1
and ^y 5:51 log cfu g
)1
and the
corresponding sample means using t tests do not give
signicant results, t = )173 and t = 110 respectively.
For the minimum value, the difference of ^y 1:69 log
cfu g
)1
and y 0:78 log cfu g
)1
neither leads to a
signicance, since t = 161. The t of the mixture model
for LAB is not as good as for clostridia, cf. Table 7.
An important reason for the lack of signicances is that
the sample means are based on only three observations.
Discussion
Increasing the proportion of Moringa elevated the CP
concentration of silage. The results (range of CP
concentrations of 84148 g kg
)1
DM) are in agreement
with other studies in that foliage from fodder trees
increases the CP concentration of silages (Ca rdenas
et al., 2003; Phiri et al., 2007). Very little WSC remained
after fermentation, with the exception of silage made
entirely with sugar cane. All treatments without sugar
cane had WSC concentrations below the detection limit
of 50 g kg
)1
DM, which is consistent with results from
McDonald et al. (2002). The NDF concentration of
the silages ranged from 396 to 775 g kg
)1
DM and
decreased when the proportion of Moringa was
increased. These results are similar to those reported
by Phiri et al. (2007) and could be attributed to the low
concentration of NDF in Moringa.
In silages made from tropical grasses, a pH value of
42 has been reported as the maximum to consider
silage to be well-preserved (McDonald et al., 2002;
Ca rdenas et al., 2003). Weissbach (1996) presented a
critical limit for good quality silage depending on DM
content in which pH should be no higher than 00257
DM content (expressed as %) + 371. With one
exception, pH of the silages where the proportion of
Elephant grass exceeded 033 was above the limit
suggested by Weissbach (1996), suggesting that Ele-
phant grass had a negative impact on silage quality.
High-quality silage is likely to be achieved when
lactic acid is the predominant acid produced, as it is
the most efcient fermentation acid and reduces silage
pH more efciently than other fermentation products.
According to the mixture experiment model, the
highest concentration of lactic acid is obtained with a
mixture of 095 Moringa and 005 molasses giving
a predicted lactic acid concentration of 105 g kg
)1
DM,
a value very close to the value actually obtained for
this silage, 106 g kg
)1
DM. This provides evidence that
Moringa silages have lactic acid concentrations supe-
rior to silage made from Elephant grass or sugar cane
and in contrast to ndings reported by Ca rdenas et al.
(2003) and Phiri et al. (2007). Higher concentrations of
fermentation acids are associated with lower DM
contents (McDonald et al., 2002; Pinho et al., 2004)
but, since the DM content was not signicantly lower
for the treatments high in Moringa when tested as a
covariate in the ANOVA ANOVA model, it is not a valid
explanation. The higher CP concentration in Moringa
should also lead to a higher buffering capacity in
mixtures with high proportions of Moringa although
Table 9 Variables modelled by the second order mixture
experiment model [concentrations of clostridia and lactic acid
bacteria (LAB)]. Estimated coefcients with standard errors (s.e.
of coefcient) for nal predicting equation y and values of
components x
1
(Moringa)
,
x
2
(Elephant grass)
,
x
3
(sugar cane)
and x
4
(molasses) giving proportions of these components at
maximal (prop. at max.) and minimal (prop. at min.) predicted
values. Corresponding sample mean y is given provided the
composition exists in Table 2.
Variable x
1
x
2
x
3
x
4
x
2
x
3
y y
Clostridia
Coefcient 345 221 331 1004 674
s.e. of coefcient 027 031 031 626 154
Prop.at max. 0 042 058 0 449 Non-
existent
Prop. at min. 0 099 001 0 228 Non-
existent
LAB
Coefcient 345 199 187 )397 1622
s.e. of coefcient 053 061 061 1224 302
Prop.at max. 0 050 050 0 598 Non-
existent
Prop. at min. 0 095 0 005 169 078
370
B. Mendieta-Araica et al.
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
this was not determined in this experiment. A high
buffering capacity also results in an elevated lactic acid
production in order to reach a satisfactory low pH
(McDonald et al., 2002).
There was a great variation in acetic acid concentra-
tion among treatments which has already been reported
for tropical silages (Ferrari and Lavezzo, 2001; Pinho
et al., 2004). Even so, the acetic acid concentration did
not reach 60 g kg
)1
DM in any of the treatments, which
is considered as the highest recommended level (Ca rd-
enas et al., 2003). Silages usually contain traces of
propionic acid and indeed the concentrations found in
the treatments studied here are within or slightly below
the range 0639 g kg
)1
DM reported by others (Pand-
itharatne et al., 1986; Lavezzo et al., 1990; Chiou et al.,
2000; Ferrari and Lavezzo, 2001; Pinho et al., 2004).
The model for estimating weight loss during the
ensiling process gave a comparatively high coefcient
for the component, sugar cane, showing clearly that
silage with sugar cane gave high weight losses. Similar
results were reported by Pedroso et al. (2005, 2008)
who pointed out that increased ethanol formation was
associated with increased weight loss.
Many microorganisms are found in fresh forages and,
under the anaerobic conditions that characterize the
ensiling process, some of the most important microor-
ganisms are LAB, clostridia, Enterobacteria and fungi.
Clostridia is not uncommon in raw material, thus, its
mere presence in feedstuffs may be unavoidable. When
clostridia are found in silage it is most likely the result of
faecal or soil contamination. Clostridia were taken into
account for their primary role in organoleptic and
nutritional decay associated with foul smell and
increased DM losses. Among the spore-forming clostri-
dia commonly found in silages, Clostridium perfringens
was chosen since it also can be a potential health hazard
for livestock (Berghaus et al., 2005). Clostridia grow
best at pH values of 7074. They cannot tolerate acid
conditions and a pH of 42 is usually considered as being
low enough to inhibit growth. In the present experi-
ment, treatments with higher pH values also had more
clostridia growth. The variables studied in the second-
order mixture experiment model were clostridia and
LAB. Concerning clostridia, a mixture of 042 of
Elephant grass and 058 of sugar cane gave the highest
counts of clostridia while 099 of Elephant grass and
001 of sugar cane gave the lowest. The means obtained
in the ANOVA ANOVA analysis also showed that the mixtures
with mainly Elephant grass had lower clostridium
values, near the minimum obtained in the model,
whereas the mixtures with Elephant grass and sugar
cane had high values near the obtained maximum
value of the mixture experiment model. In treatments
containing sugar cane, cultivation and harvest methods
lead to a high probability of soil contamination which
could be the main reason for the high clostridia levels
among these treatments.
In order to preserve high-moisture forages such as
Moringa, a high LAB activity is necessary. Therefore,
the rapid establishment and consequent maintenance
of anaerobic conditions throughout ensiling is essential
for LAB proliferation. For silages made from temperate
grasses, LAB levels of at least 39 log cfu g
)1
are
desirable for a good fermentation process but lower
values have been reported as normal in silages from
tropical grasses (Tjandraatmadja et al., 1994a,b; Pedroso
et al., 2005). The low concentration of lactic acid
combined with high LAB counts in treatment
M0E67Sc33 is unusual but could indicate that there
was an unintentional minor air leakage leading to the
presence of aerobic fungi that consumed the lactic acid
as it was produced. In the second-order mixture
experiment model used in the present experiment, a
mixture of 050 of Elephant grass and 050 of sugar cane
gave the highest LAB count. This is not quite in line
with the highest lactic acid concentration, produced in
the mixture of 095 of Moringa and 001 of Molasses,
demonstrating that counts of viable bacteria and lactic
acid concentrations do not always harmonize.
When exposed to oxygen in the air, all silages
deteriorate as a result of aerobic microbial activity. As
soon as a silo is opened, it is exposed to the air for the
entire feed-out period. Therefore, silage quality is not
only characterized by its chemical composition, but also
by how long it will remain stable and resist deteriora-
tion once the silo has been opened and the contents are
fed to livestock. The start of the deterioration process is
characterized by a rise in temperature in the silage
indicating increased microbial activity as the silage
spoils.
Treatment M66E33Sm1 had an unexpectedly long
time to spoilage and M63E32Sm5 unexpectedly short,
a difference which was signicant in the ANOVA ANOVA
analysis. When the rst-order mixture experiment
model was used for CO
2
production during the storage
stability phase the mixture with the highest CO
2
production was 095 of Moringa and 005 of molasses.
The model suggested 099 of Elephant grass and 001
of sugar cane would have the lowest CO
2
production.
This indicates that silage with Moringa could have a
somewhat lower aerobic stability after opening the
silo. Nevertheless, there was a great variation in time
to spoilage values observed in this study, something
that has also been presented in other reports under
tropical conditions (Danner et al., 2003; Pedroso et al.,
2008). A relationship between acetic acid and stability
was proposed by Danner et al. (2003) who claimed
that increasing acetic acid concentrations inhibit spoil-
age organisms, thereby promoting exponential
increases in stability.
Silages made from tropical grasses and Moringa
371
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
Conclusion
The CP concentration of silage based on Elephant grass
or sugar cane was substantially increased by adding
Moringa. The present study shows that the inclusion of
Moringa also has a positive inuence on silage quality.
The pH is lowered and the production of lactic acid is
elevated. Furthermore, ensiling mixtures of Moringa
and sugar cane decreased DM losses compared with
silage based on only sugar cane. By using the mixture
experiment model when evaluating the result of the
fourteen treatments, modelling of other proportions
than those used in the experiment could be performed.
Such operation indicates that, even if ensiling Moringa
alone would give the best fermentation pattern, the
aerobic stability would be better when Moringa is
ensiled in mixtures with Elephant grass or sugar cane.
Equal proportions of Moringa and Elephant grass with
005 molasses added or equal proportions of Moringa
and sugar cane would provide both acceptable fermen-
tation patterns and stability after opening while still
maintaining the nutritive value of the silage. The
experiment also showed that Moringa alone with 001
005 molasses added produces a good silage quality.
Acknowledgments
The funding of this research by the Swedish Interna-
tional Development Cooperation Agency (SIDA) is
gratefully acknowledged.
References
ANDRADE NDRADE S. and MELOTTI ELOTTI L. (2004) Effect of some
additives on the quality of elephant grass (Pennisetum
purpureum, Schum) silage. Brazilian Journal of Veterinary
Research and Animal Science, 41, 409415.
AOAC (2000) Ofcial methods of analysis. Washington, DC,
USA: Association of Ofcial Analytical Chemists.
APHA (2001) Compendium of methods for the microbiological
examination of foods, 4th edn. Washington, DC, USA:
American Public Health Association.
ASHBELL SHBELL G., WEI NBERG EI NBERG A., AZRI ELI ZRI ELI A., HEN EN Y. and ORBE RBE
B. (1990) A simple system to study the aerobic
determination of silages. Canadian Agricultural
Engineering, 33, 391393.
ATKI NSON TKI NSON A.C. and DONEV ONEV A.N. (1992) Optimum experi-
mental designs. Oxford, UK: Oxford University Press.
BERGHAUS ERGHAUS R., MCCLUSKEY CCLUSKEY B. and CALLAN ALLAN R. (2005) Risk
factors associated with hemorrhagic bowel syndrome in
dairy cattle. Journal of the American Veterinary Medical
Association, 226, 17001706.
CARDENAS A RDENAS J., SANDOVAL ANDOVAL C. and SOLORI O OLORI O F. (2003)
Chemical composition of grass and forage trees mixed
silages. Tecnica Pecuaria en Mexico, 41, 283294.
CHI OU HI OU P., CHANG HANG S. and YU U B. (2000) The effects of wet
sorghum distillers grains inclusion on Napier grass silage
quality. Journal of the Science of Food and Agriculture, 80,
11991205.
DANNER ANNER H., HOLZER OLZER M., MARYHUBER ARYHUBER E. and BRAUN RAUN R.
(2003) Acetic acid increases stability of silage under
aerobic conditions. Applied and Environmental
Microbiology, 69, 562567.
FERRARI ERRARI E. and LAVEZZO AVEZZO W. (2001) Quality of Elephant
grass silage (Pennisetum purpureum Schum) wilted or
adding cassava meal. Scientia Agricola, 30, 14241431.
LAVEZZO AVEZZO W., LAVEZZO AVEZZO O. and BONASSI ONASSI I. (1990) Efeito do
emurcheicimento, formol, a cido fo rmico e soluca o de
Viher sobre a qualidade de silagens de capim-Elefante,
cultivares Mineiro e Vruckwona (Effect of wilting,
formol, formic acid and Vilher solution on silage
quality of Elephant grass, cv. Mineiro and Vruckwona).
Pesquisa Agropecua ria Brasileira, 25, 125134.
MAKKAR AKKAR H. and BECKER ECKER K. (1996) Nutritional value and
anti-nutritional components of whole and ethanol
extracted Moringa oleifera leaves. Animal Feed Science and
Technology, 63, 211228.
MAKKAR AKKAR H. and BECKER ECKER K. (1997) Nutrients and anti-
quality factors in different morphological parts of the
Moringa oleifera tree. Journal of Agricultural Science,
Cambridge, 128, 311322.
MC CDONALD ONALD P., HENDERSON ENDERSON N. and HERON ERON S. (2002) The
biochemistry of silage. Marlow, UK: Chalcombe Publications.
MC CDOWELL OWELL R. (1972) Improvement of livestock production in
warm climates. San Francisco, CA, USA: Freeman.
MTENGETI TENGETI E., KAVANA AVANA P., URI O RI O N. and SHEM HEM M. (2006)
Chemical composition and fermentative quality of fodder
grasses ensiled with de-rinded fresh sugarcane crush.
Tropical and Subtropical Agroecosystems, 6, 157165.
PANDI THARATNE ANDI THARATNE S., ALLEN LLEN V., FONTENOT ONTENOT J. and JAY- AY-
ASURI YA ASURI YA M. (1986) Ensiling characteristics of tropical
grasses as inuenced by stage of growth, additives
and chopping length. Journal of Animal Science, 63,
197207.
PEDROSO EDROSO A., NUSSI O USSI O L., PAZI ANI AZI ANI S., SANTANA ANTANA D., IGARASI GARASI
M., COELHO OELHO R., PACKER ACKER H., HORI I ORI I J. and GOMES OMES L.
(2005) Fermentation and epiphytic microora dynamics
in sugar cane silage. Scientia Agricola, 62, 427432.
PEDROSO EDROSO A., NUSSI O USSI O L., LOURES OURES D., PAZI ANI AZI ANI S., RI BEI RO I BEI RO
J., MARI ARI L., ZOPOLLATTO OPOLLATTO M., SCHMI DT CHMI DT P., MATTOS ATTOS W.
and HORI I ORI I J. (2008) Fermentation, losses and aerobic
stability of sugarcane silages treated with chemical or
bacterial additives. Scientia Agricola, 65, 589594.
PHI RI HI RI M., NGONGONI GONGONI N., MAASDORP AASDORP B., TI TTERTON I TTERTON M.,
MUPANGWA UPANGWA J. and SEBETA EBETA A. (2007) Ensiling
characteristics and feeding value of silage made from
browse tree legume-maize mixtures. Tropical and
Subtropical Agroecosystems, 7, 149156.
PI NHO I NHO E., COSTA OSTA C., ARRI GONI RRI GONI M., SI LVEI RA I LVEI RA A.,
PADOVANI ADOVANI C. and PI NHO I NHO S. (2004) Fermentation and
nutritive value of silage and hay made from the aerial
part of cassava (Manihot sculenta Crantz). Scientia Agricola,
61, 364370.
REYES EYES-SANCHEZ A NCHEZ N., LEDI N EDI N S. and LEDI N EDI N I. (2006) Biomass
production and chemical composition of Moringa oleifera
under different management regimes in Nicaragua.
Agroforestry Systems, 66, 231242.
372
B. Mendieta-Araica et al.
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
SAS (2004) SAS STAT Users Guide Version 9.1. Cary, NC,
USA: SAS Institute Inc.
TJ ANDRAATMADJA JANDRAATMADJ A M., NORTON ORTON B. and MACRAE ACRAE I.
(1994a) Ensilage characteristics of three tropical grasses
as inuenced by stage of growth and addition of
molasses. World Journal of Microbiology and Biotechnology,
10, 7481.
TJ ANDRAATMADJA JANDRAATMADJ A M., NORTON ORTON B. and MACRAE ACRAE I.
(1994b) Ensilage of tropical grasses mixed with legumes
and molasses. World Journal of Microbiology and
Biotechnology, 10, 8287.
UMANA MAN A R., STAPLES TAPLES C., BATES ATES D., WI LCOX I LCOX C. and
MAHANNA AHANNA W. (1991) Effects of a microbial inoculant
and (or) sugar cane molasses on the fermentation,
aerobic stability and digestibility of Bermuda grass
ensiled at two moisture contents. Journal of Animal
Science, 69, 45884601.
VAN AN SOEST OEST P.J., ROBERTSON OBERTSON J.B. and LEWI S EWI S B.A. (1991)
Methods for dietary bre, neutral-detergent bre and
non-starch polysaccharides in relation to animal
nutrition. Journal of Dairy Science, 4, 35833597.
WEI SSBACH EI SSBACH F. (1996) New developments in crop conservation.
XI International Silage Conference. IGER, Aberystwyth,
UK, pp. 1125. Abertsywyth, UK: IGER.
ZANI NE ANI NE A., SANTOS ANTOS E., FERREI RA ERREI RA D., OLI VEI RA LI VEI RA J.,
ALMEI DA LMEI DA J. and PEREI RA EREI RA O. (2006) Evaluation of silage
of Elephant-grass with addition of wheat meal. Archives
de Zootechnie, 55, 7584.
Silages made from tropical grasses and Moringa
373
2009 Blackwell Publishing Ltd. Grass and Forage Science, 64, 364373
V
ORIGINAL RESEARCH
Feeding Moringa oleifera fresh or ensiled to dairy
cowseffects on milk yield and milk flavor
Bryan Mendieta-Araica & Eva Sprndly &
Nadir Reyes-Snchez & Rolf Sprndly
Accepted: 10 February 2011 / Published online: 23 February 2011
# Springer Science+Business Media B.V. 2011
Abstract Moringa oleifera, either fresh or ensiled, was
compared with Elephant grass as a main feedstuff for dairy
cows. To test the effects feed had on milk yield, milk
composition, ration digestibility, and the organoleptic charac-
teristics of milk, six lactating dairy cows were used in a
Changeover 33 Latin Square experiment, replicated twice.
With equal intake of metabolizable energy the intake of
protein and fiber differed (p<0.001) between all diets where
fresh Moringa had the highest and the Elephant grass diet
had the lowest intake. Compared with the control diet,
ensiled Moringa had higher digestibility (P<0.05) of both
protein and fiber. With the exception of DM digestibility, no
digestibility differences were found between fresh Moringa
and Moringa silage treatments. Milk yield did not differ
between any of the treatments and averaged 13.7 kg
cow day
1
. Milk composition was similar among all treat-
ments. Milk from the fresh Moringa treatment, however, had
a grassy flavor and aroma, significantly different from the
other two treatments, even though it was normal in color and
appearance. No organoleptic differences were found between
milk from the control treatment and the Moringa silage
treatment. The conclusion is that Moringa silage can be fed to
dairy cows in large quantities to produce the same quantity
and quality of milk as traditional diets.
Keywords Fresh Moringa
.
Silage
.
Dairy cows
.
Milk yield
.
Organoleptic characteristics
Introduction
Livestock production is a very important part of the
agricultural sector in many tropical countries, representing
up to 40% of the agricultural gross domestic product
(Steinfeld et al. 2006). The dairy industry, in particular, is
of increasing economic importance. For example, milk
production in Latin America increased by 12% between
2003 and 2007 (FAO 2010).
Although regional variation occurs throughout Central
and South America, it is fair to describe intensive dairy
farming in Latin America as consisting of medium- to
small-scale farms with 1017 ha and a herd of 1521 cows
(Holmann et al. 2003). These farms generally practice stall-
feeding, providing mainly planted Elephant grass (Pennisetum
purpureum) as roughage and either molasses, locally avail-
able by-products, or commercial feeds as concentrates. Pure
HolsteinFriesian cattle are the most popular and produce
approximately 16 kg milk per day (range, 527 kg) at
concentrate feeding rates of 110 kg per day (de Leeuw et al.
1998). Even though such roughage has a low nutritional
content, few farmers can afford to sufficiently supplement the
diet with conventional concentrates during the dry season
because they cost too much.
Despite the enormous potential for biomass production
in tropical regions, forages are only abundant during the
rainy season and many authors (Ser et al. 1995; Holmann
et al. 2003; Ruiz 2005) agree that the main restriction to
B. Mendieta-Araica
:
E. Sprndly
Department of Animal Nutrition and Management,
Swedish University of Agricultural Sciences,
P.O. Box 7024, SE-750 07 Uppsala, Sweden
N. Reyes-Snchez
Facultad de Ciencia Animal, Universidad Nacional Agraria,
P.O. Box 453, Managua, Nicaragua
R. Sprndly (*)
Department of Animal Nutrition and Management,
Swedish University of Agricultural Sciences,
Kungngen Research Station, SE-753 23 Uppsala, Sweden
e-mail: rolf.sporndly@huv.slu.se
Trop Anim Health Prod (2011) 43:10391047
DOI 10.1007/s11250-011-9803-7
increased milk production is the low quality and quantity of
feed resources during the dry season.
Grass silages can be used to overcome feed shortage
during the dry season; however, tropical grasses mature
quickly and produce low quality silages due to the low
nutritive value of the biomass ensiled (Panditharatne et al.
1986; Tjandraatmadja et al. 1994). The problem remains
to find inexpensive solutions to the nutritional deficit
created by scarcity of good quality forages during the dry
season.
The potential for agroforestry to supply animals with
high quality feed has been recognized by many authors
(Pezo 1991; Fernndez-Baca 1992; Szott et al. 2000;
Mauricio et al. 2008). Trees and bushes are drought
tolerant and the foliage is often rich in crude protein (CP)
(Benavides 1994; Crdenas et al. 2003). Furthermore, they
often show a tolerance to a wide range of management
practices (Paterson et al. 1998) and may thus enhance the
sustainability of the farming system (Reyes-Snchez
2006).
One of the most interesting trees is Moringa oleifera,
commonly referred to as Moringa or Drumstick tree,
which grows throughout the tropics. It is one of the most
widely used species for fodder (Makkar and Becker 1996).
It can grow in all types of soils (Duke 1983) and can
tolerate dry seasons lasting up to 6 months. The total dry
matter (DM) yield from Moringa can be up to
24 Mg ha
1
year
1
(Reyes-Snchez et al. 2006a). Further-
more, Moringa contains negligible amounts of antinutri-
tional factors, has a high CP content and contains
significant amounts of vitamins A, B, and C in the foliage
(Makkar and Becker 1996; Ferreira et al. 2008).
As fresh forage, Moringa has been included into the
diets of many different animals. Positive effects on the
feeding behavior in goats (Manh et al. 2005) and on the
growth rate in sheep (Ben Salem and Makkar 2009) have
been reported. Fresh Moringa in dairy cow diets has also
given favorable production results (Reyes-Snchez et al.
2006b) as has Moringa leaf meal as a protein source in the
concentrate (Mendieta-Araica et al. 2010). Studies have
also shown that Moringa can be ensiled alone or in
mixtures with Elephant grass or sugar cane to increase the
nutritive value of the silage (Mendieta-Araica et al. 2009).
However, diets based almost entirely on Moringa or
Moringa silage have not been previously studied in detail
for dairy cows.
Therefore, the aim of this study was to evaluate the milk
yield, milk composition and digestibility of three different
dairy cow diets: (1) a conventional Elephant grass diet with
commercial concentrate, (2) fresh Moringa only, or (3)
ensiled Moringa only. Special attention was even paid to
the organoleptic characteristics of milk produced by the
cows on these diets.
Materials and methods
Experimental design
The experiment was designed as a Changeover 33 Latin
Square, replicated in two orthogonal Latin squares, as
described by Patterson and Lucas (1962). Each experimental
period consisted of 2 weeks for treatment adaptation and
2 weeks of data collection with regard to milk yield and feed
intake. The last week of each period was used for
organoleptic testing and to estimate digestibility.
Location
The experiment was carried out during the dry season at the
National Agrarian University (UNA) farm in Managua,
Nicaragua. The farm is located 12 08 15 N and 86 09 36
W. The average annual temperature is 27C and the mean
annual rainfall is 1,440 mm, with a marked dry season between
November and May.
Feed preparation
The fields with Elephant grass (P. purpureum cv CT 115)
and M. oleifera were already divided into regularly
harvested plots before the start of the experiment in order
to enable circa 45-day harvest intervals throughout the
experiment. Each day, Elephant grass and Moringa (leaves
and soft twigs) from a new plot was harvested, chopped
mechanically into 2 cm lengths and offered fresh to the
cows, resulting in re-growth intervals of 463 days.
Moringa silage was prepared 120 days before the experi-
ment by cutting a field of M. oleifera 45 days after a
previous harvest. After removing the thicker twigs, the
material was chopped and sugar-cane molasses was added
to at a rate of 5% according to fresh weight. The material
was then put into 5597 cm polythene bags and com-
pressed by hand. The bags were then sealed to serve as
silos. In total 180 bags, weighing approximately 45 kg
each, were stored indoors on a concrete floor until opening.
Feed concentrate was mixed at the UNA Feed Concentrate
Plant using commercially available feedstuffs: 41% rice
polishing, 30% sorghum, 20% soybean meal, 6% molasses,
1.5% calcium carbonate, 1% peanut meal, and 0.5%
sodium chloride.
Management
Six dairy cows of the Brown Swiss breed from the farm
herd, each in their second or third lactation and weighing
on average 456 kg (SD 50), were used in the trial. The
cows were in their fourth week of lactation at the start of
the experiment. Prior to the experiment, they were injected
1040 Trop Anim Health Prod (2011) 43:10391047
with vitamin A (625,000 IU), vitamin D
3
(125,000 IU), and
vitamin E (125 IU), treated against external and internal
parasites and vaccinated against anthrax. All of the cows
were treated according to EC Directive 86/609/EEC for
animal experiments.
The cows were kept loose in individual, well-ventilated
stalls with concrete floors. Water and minerals were supplied
ad libitum. Except during the weeks when fecal samples were
taken, the cows were allowed to exercise daily in a common
area while the individual stalls were cleaned.
Treatments
The cows were fed three different isocaloric treatments
during three periods. The diets in these treatments are
presented in Table 1. They were designed to fulfill DM and
metabolizable energy (ME) requirements (NRC 1988). The
expected DM intake was calculated with the NRC (2001)
equation using the average body weight and milk yield for
the entire herd at the start of the experiment. The control
treatment (60% Elephant grass roughage and 40% com-
mercial concentrate) represents the current recommendation
for dairy farmers in Central America (Vlez 1997 and
Castro Ramrez 2002). The Moringa treatments contained
only Moringa with 1 kg added molasses to promote
palatability.
Roughages were offered twice daily to each cow at 0700
and 1700 hours, while the concentrate of the control
treatment was fed during milking, at 0500 and 1600 hours.
The Moringa treatments contained the supplement molasses
which was thoroughly mixed with the roughages before
feeding. The DM content of roughages was determined
twice a week using a microwave oven according to the
procedure described by Undersander et al. (1993) so that
the DM allowance could be adjusted according to the
feeding plan.
The amounts of feed offered were weighed daily and
sampled. Sampling was performed as follows: 1 kg of offered
roughage per cow per day was collected and immediately
frozen at 18C. The frozen samples from each period were
then thawed and pooled into one sample for chemical analysis.
The occasional refusals were individually weighed and frozen
samples were sent for chemical analysis at the end of each
experimental period. One kilogram of concentrate was
collected as a sample every week during data collection for
chemical analysis (see below).
The cows were hand-milked twice daily with approxi-
mately 12-h interval. At each milking, the yield from each
cow was weighed and recorded, and 100-ml samples were
taken in sterile vials. The milk samples were immediately
refrigerated at 4C and then pooled into one sample per
cow per data collection period for fat and protein content
analysis. The same milk sampling procedure was repeated
during the last 3 days of each experimental period for
organoleptic testing.
Digestibility study
The cows were weighed before the start of the experiment
and after each period. During the last week of each period,
rubber mats were placed in the stalls while all of the feces
from each cow were manually collected. During those
weeks, the cows were supervised around the clock.
Whenever a cow adopted the defecation position, a shovel
was put under her tail to collect the feces, thereby
minimizing urine and dirt contamination. The feces from
each cow were put into a large container and covered with a
lid to avoid evaporation. Once daily, the feces from each
container were weighed and thoroughly mixed. Five
percent were taken as a subsample and frozen before the
containers were emptied. When the collection was com-
plete, the subsamples from each cow were thawed and
mixed together. Approximately 300 g of this mixture was
then taken as a fecal sample for each animal and
experimental period to be used in the chemical analysis.
Chemical analysis and organoleptic tests
The feed samples and fecal samples were analyzed so as to
be able to estimate apparent digestibility. The samples were
dried and ground through a 1-mm sieve before analysis.
Ash and DM content were analyzed using AOAC (1990)
procedures. CP content was determined using the Kjeldahl
method (AOAC 1984). Neutral detergent fiber (NDF) and
acid detergent fiber (ADF) contents were analyzed using
sodium sulfite as described by Van Soest et al. (1991). An
apparent digestibility coefficient for DM was calculated by
comparing the dietary intake with recovery in the feces.
Organic matter digestibility of the roughages was determined
using in vitro incubation in rumen liquid. To determine ME,
0.5-g dried sample was incubated with 49-ml buffer and 1-ml
rumen fluid for 96 h at 38C. Thereafter, the residues were
weighed and combusted and the organic matter digestibility
Table 1 Feeding treatments offered to dairy cows in a Changeover
33 Latin Square experiment, replicated twice
Treatment Control Fresh Moringa Moringa silage
Feed (kg DM day
1
)
Elephant grass 6.65
Moringa foliage 10.31
Moringa silage 10.40
Molasses 1.00 1.00
Concentrate 4.44
Trop Anim Health Prod (2011) 43:10391047 1041
coefficient could be determined. ME was estimated using the
equation presented by Lindgren (1979), except for in the
concentrate where the analysis was provided by the feed
producer (UNA Feed Concentrate Plant, Nicaragua).
Nitrogen content in the milk was determined using the
Kjeldahl method and milk protein content was calculated as
N6.38. The Babcock method used to determine milk fat
content and the methods used to analyze total solids and
casein were all according to AOAC (1984) procedures.
An experienced panel evaluated the organoleptic char-
acteristics of the milk samples. A triangle difference test
(Witting de Penna 1981) was applied using a milk sample
with normal sensory characteristics (flavor, aroma, color,
and appearance) as standard. Twenty judges were asked to
determine the sensory characteristics in the milk samples
according to the score sheet presented in Table 2, where 5 is
the maximum score for flavor and aroma and 3 the
maximum for color and appearance. The total score for
each trait is the sum of scores awarded by all of the judges,
while classification is the most common score awarded.
Statistical analysis
The data were analyzed using the GLM procedure in SAS
Software Version 9.1.2 (SAS 2004). Tukey's pairwise
comparison was used whenever the overall F test of
treatment means showed a significant result. The mathe-
matical model used for digestibility, milk production, and
organoleptic data was the following:
Y
ijkl
= + B
i
+P
j
+T
k
+C
l
+
ijkl
, with i =1,2, j =1,,6 and
k=1,2,3, where was the overall mean, B
i
the fixed effect
of block, P
j
the fixed effect of period, T
k
the fixed effect of
treatment, C
l
the random effect of cow, and
ijkl
the random
residual error.
Carry-over effects from previous periods and interaction
between period and treatment were tested, and then excluded
from the final model as they were not significant (P>0.10).
Results
Nutrient intake and apparent digestibility
The chemical compositions of the feeds used in the three
treatments are presented in Table 3. Both fresh Moringa and
Moringa silage had high CP contents and low NDF
concentrations compared with the Elephant grass used in
the experiment. The CP content in the concentrate was
typical for commercial concentrates in Central America.
Intake of DM was approximately as planned (Tables 1
and 4) as there were no refusals in the control treatment and
feed refusals in the Moringa treatments were minor,
representing only 0.97% of the daily offered amount. There
were no significant differences between treatments in ME
intake, showing that the treatments were isocaloric as
planned. There was also no difference in the live weight
change between the treatments (data not shown) but a slight
decline in live weight over the complete experiment period
(0.19 kg day
1
) indicated an underfeeding equivalent to
6.5 MJ ME day
1
(NRC 1988).
Nutrient intake and apparent digestibility coefficients
among the three treatments are presented in Table 4. There
Table 2 Descriptive terms and scores for testing the organoleptic characteristics of milk (adapted and shortened from Witting de Penna 1981)
Aroma Score Flavor Score Color Score Appearance Score
Undoubted characteristic 5 Undoubted characteristic 5 Yellowish white 3 Homogeneous when shaken 3
Normal 4 Normal 4 Markedly white 2 Small lumps after shaking 2
Subtly lack of freshness 3 Subtly impure 3 Abnormal coloration 1 Big lumps after shaking 1
Subtly grassy 2 Subtly grassy 2
Markedly impure 1 Markedly grassy 1
Constituents Elephant grass Fresh Moringa Moringa silage Molasses Concentrate
DM (g kg
1
) 153(17.3) 193(1.2) 267(9.8) 727(0.1) 845(10.4)
CP (g kg
1
DM) 108(2.0) 241(5.2) 226(5.7) 22(0.3) 184(1.6)
NDF (g kg
1
DM) 510(7.7) 453(5.8) 435(9.1) nd 91(3.1)
ADF (g kg
1
DM) 314(8.1) 299(5.3) 291(4.0) nd 58(1.1)
Lignin (g kg
1
DM) 30(5.8) 103(6.1) 99(3.3) nd 43(4.5)
Ash (g kg
1
DM) 181(0.3) 93(3.2) 116(3.2) 28(0.1) 77(4.0)
ME (MJ kg
1
DM) 8.5(0.08) 10.9(1.50) 10.8(1.72) 9.2(0.00) 14.6(0.12)
Number of samples 6 6 6 6 6
Table 3 The chemical
composition of feedstuffs
fed to dairy cows presented
as means with standard
deviations in parenthesis
DM dry matter, CP crude
protein, NDF neutral detergent
fiber, ADF acid detergent
fiber, ME metabolizable
energy, MJ megajoule,
nd not determined
1042 Trop Anim Health Prod (2011) 43:10391047
were significant differences among treatments with regard to
all nutrient intakes with the exception of ME. Significant
differences (P<0.05) were also found for DM, CP, organic
matter, NDF and ADF digestibility. Compared with the
control diet, the Moringa silage diet had a significantly
higher digestibility of all the measured variables, while the
fresh Moringa diet only had a higher digestibility of CP and
NDF. With the exception of DM digestibility, no differences
in digestibility were found between Moringa treatments.
Milk yield and composition
The average daily milk yield during the experiment was
13.7 kg. There were no significant differences between
treatments with regard to yield or milk composition. On
average, the milk contained 35.1 g fat kg
1
, 123 g total
solids kg
1
, 34.5 g protein kg
1
protein and 27.3 g
casein kg
1
(Table 5).
Organoleptic characteristics
Results from the organoleptic characteristic tests are
presented in Table 6 and can be compared with the scoring
system in Table 2. The color and appearance of milk from
all three treatments were classified as normal. No differ-
ences in these traits were found among treatments.
However, there were significant differences (P<0.001) in
the scores for flavor and aroma between treatments. The
milk from the control and Moringa silage treatments were
classified as normal by most judges, while milk from the
Table 4 Least square means of nutrient intake and apparent digestibility of three different feeding treatments fed to dairy cows
Items Treatments Standard error Significance level
Elephant grass+concentrate Fresh Moringa Moringa silage
Nutrient intake (kg day
1
)
Dry matter 11.1
b
11.2
ab
11.3
a
0.03 *
Organic matter 9.6
c
10.2
a
10.1
b
0.03 **
Crude protein 1.53
c
2.48
a
2.39
b
0.01 **
Neutral detergent fiber 3.79
c
4.64
a
4.49
b
0.02 **
Acid detergent fiber 2.35
c
3.06
a
3.00
b
0.01 **
Lignin 0.39
c
1.06
a
1.01
b
0.01 **
Metabolizable energy (MJ day
1
) 121 120 120 0.41 ns
Apparent digestibility coefficients
Dry matter 0.64
b
0.69
b
0.76
a
0.02 *
Crude protein 0.74
b
0.81
a
0.83
a
0.01 *
Organic matter 0.64
b
0.70
ab
0.77
a
0.02 *
Neutral detergent fiber 0.37
b
0.54
a
0.66
a
0.04 *
Acid detergent fiber 0.33
b
0.43
ab
0.61
a
0.04 *
ns not significant
abc
Within a row means without common superscript differs
*P<0.05, **P<0.001
Table 5 Least square means of milk production and milk composition from cows fed three different treatments
Items Treatments Standard error Significance level
Elephant grass+concentrate Fresh Moringa Moringa silage
Milk (kg cow
1
day
1
) 13.9 13.6 13.7 0.13 ns
Energy corrected milk (kg cow
1
day
1
) 13.0 12.9 12.8 0.13 ns
Milk fat (g kg
1
milk) 34.9 35.3 35.1 0.17 ns
Total solids (g kg
1
milk) 122 123 123 0.37 ns
Milk crude protein (g kg
1
milk) 34.5 34.6 34.2 0.19 ns
Casein (g kg
1
milk) 27.3 27.3 27.3 0.17 ns
ns not significant
Trop Anim Health Prod (2011) 43:10391047 1043
fresh Moringa treatment was classified as subtly grassy by
most judges (Tables 2 and 6).
Discussion
Latin American farmers commonly feed Elephant grass as the
basal diet to dairy cows. However, Elephant grass matures
rapidly, losing protein content, thus resulting in the need to
supplement the basal diet with commercial concentrates. An
interesting feeding alternative is to replace Elephant grass with
Moringa in dairy cow diets. Previous studies have included
Moringa in the roughage to reduce the need for other protein
supplements (Reyes-Snchez et al. 2006b). In the present
experiment, M. oleifera constitutes the entire roughage diet
without any protein supplement and only complemented
with 1 kg of molasses. This is a level of Moringa in dairy
cow diets that to our knowledge, has not previously been
tested under experimental conditions.
Elephant grass can be fed either fresh or ensiled. In order
to ensure high quality, the Elephant grass must be cut at
regular intervals. Prolonged harvest intervals result in
dramatic decreases in both digestibility and CP content. It
is increasingly common that Latin American farmers ensile
Elephant grass at the right stage of development to ensure
quality (Reyes-Snchez et al. 2008, 2009). The nutrient
content of Moringa is less sensitive to cutting intervals and
retains better digestibility and CP content over longer
periods (Reyes-Snchez et al. 2006a). Moringa can also
be ensiled, either alone or in combination with other crops
(Mendieta-Araica et al. 2009). Although it is not as
necessary to ensile Moringa as it is Elephant Grass, it is
practical to feed Moringa from a silo rather than harvesting
and transporting the roughage daily.
CT-115, an improved variety of Elephant Grass devel-
oped by the Cuban Institute of Animal Science through
tissue culture, was used in the present experiment. The
mean CP content (108 gkg
1
DM) is within the range
(71143 gkg
1
DM) reported by other researchers for this
variety (Valenciaga et al. 2001, 2009) but higher than the
range (4982 gkg
1
DM) reported for other Elephant grass
varieties (Mendieta-Araica et al. 2009; Sarwatt et al. 2004).
The CP contents of fresh and ensiled Moringa reported in
Table 3 are considerably higher than the 144 gkg
1
DM
reported previously (Mendieta-Araica et al. 2009). This is
explained by the fact that only the leaves and soft twigs were
used in this experiment, whereas leaves, twigs, and branches
were used in the previous study. A similar pattern is presented
by Fujihara et al. (2005) who report 265 g CP kg
1
DM in
leaves compared with 195 g CP kg
1
DM in a mixture of soft
twigs and leaves. Makkar and Becker (1997) also report a
great variability in CP content for different parts of Moringa:
leaves contain 264 g CP kg
1
DM, while twigs and stems
contain 72 and 62 g CP kg
1
DM, respectively.
The main objective of this study was to see if Moringa
could maintain the same level of production as the
traditional diet of Elephant grass and concentrates, thus
enabling farmers to produce high-quality feed for their
dairy cows without having to purchase concentrates.
Moringa has a high CP content and as the diets were
designed to cover the energy requirements of the cows the
Moringa diets contained approximately 60% more protein
than the control diet. This excess protein was highly
digestible as can be seen from the digestibility values
(Table 4). The higher protein feeding did not lead to a
higher milk production and the excess nitrogen had to be
excreted in the urine at an energy cost for the animal. The
higher digestibility of the nutrients in the Moringa treat-
ments could theoretically provide this extra energy.
The digestibility of Moringa silage has not previously
been studied; however, the digestibility values observed
were better than for silages from others forage trees and
shrubs. Crdenas et al. (2003) reports DM in vitro
digestibility values for four forage tree silages between
Table 6 Organoleptic evaluation scores and average sensory classification for milk from cows fed one of three treatments
Attribute Max score Elephant grass+concentrate Fresh Moringa Moringa silage Significance level
Total score Classification Total score Classification Total score Classification
Flavor 100 79
a
4 45
b
2 81
a
4 *
Aroma 100 80
a
4 47
b
2 79
a
4 *
Color 60 60 3 60 3 59 3 ns
Appearance 60 59 3 59 3 59 3 ns
Total scores are the sum of the scores given by 20 experienced judges
High scores correspond to high milk quality
ns not significant
ab
Scores within a row without common superscripts differ significantly
*P<0.05
1044 Trop Anim Health Prod (2011) 43:10391047
45% and 58% and organic matter in vitro digestibility for
the same silages between 46% and 59%. In a study with
mixed silages, Acacia boliviana with maize had a DM
apparent digestibility of 53.5% and Leucaena leucocephala
with maize 56.5% (Phiri et al. 2007).
As mentioned earlier, there were no significant differ-
ences in milk production between treatments despite
differences in CP intake at a production level of
13.7 kg cow day
1
for milk yield, or 12.9 kg cow day
1
ECM. According to Oldham (1984), lactating cows fed
diets which meet their ME and CP requirements are not
expected to respond to increased protein levels but a
reduced ME utilization can sometimes be observed with
excessive CP intakes. A higher CP intake, however, is often
associated with a higher dry matter intake (Kokkonen et al.
2002). Restricted feed allowance, as practiced in the present
experiment, did not allow a higher feed intake and this
aspect could therefore not be studied.
Milk composition was not affected by the treatments,
which is consistent with previous findings that show no
relationship between milk protein content and CP content in
the diet when the diets have similar energy concentrations
(Sutton 1989; Schingoethe 1996; Reyes-Snchez et al.
2006a).
In countries where farmers feed Moringa to dairy cows,
there are reports that it gives an off-flavor to the milk
(Reyes-Snchez 2006). To avoid a grassy flavor and aroma
in the milk, Agrodesierto (2010) recommends that during at
least 3 h prior to milking, fresh Moringa should not be fed
to cows. This recommendation is, however, difficult for
intensive dairy farmers to follow. In a study by Reyes-
Snchez et al. (2006a) where 20% of the diet consisted of
Moringa, no effects in milk flavor were observed. In the
present experiment, the cows fed fresh Moringa with
molasses produced milk with inferior organolpetic charac-
teristics. It is important to note that we found that the same
quantity of ensiled Moringa with molasses did not
negatively affect milk quality.
How flavor is passed on to milk has been widely studied
(Dougherty et al. 1962; Shipe et al. 1962; Shipe et al.
1978). Makkar and Becker (1997) attribute the off-taste in
milk from cows fed fresh Moringa to glucosinolates. Shipe
et al. (1962) report that the presence of esters gives off-taste
to milk when introduced into either the rumen or the lungs.
Accordingly, Bennet et al. (2003) report appreciable
concentrations of 4-(-Lrhamnopyranosyloxy)-bensylglu-
cosinolate, three monoacetyl isomers of this glucosinolate
and the esters 3-caffeoylquinic and 5-caffeoylquinic in
Moringa leaves.
Ensiling results in up to a 90% reduction of glucosino-
lates concentrations (Nash 1985; Fales et al. 1987; Panciera
et al. 2003). That Moringa silage did not lead to the same
off-flavor and aroma in milk as fresh Moringa did could be
due to a reduction in the concentration of glucosinolates
already mentioned.
In conclusion, the results of this study show that
ensiled Moringa can be fed to dairy cows in large
quantities without any negative effect on nutrient intake
or digestibility. Cows fed large quantities of Moringa
produce milk in as much quantity and as high in quality
as cows fed conventional Elephant grass diets. Whereas a
fresh Moringa diet can lead to an off-flavor and aroma in
milk, a Moringa silage diet leads to good organoleptic
milk characteristics.
Acknowledgments The funding for this research, provided by the
Swedish International Development Agency (Sida), is gratefully
acknowledged. Brje Ericsson, is also acknowledged for his help
and support throughout the laboratory analysis.
References
Agrodesierto, 2010. MoringaMoringa oleifera, Programas agrofores-
tales, Retrieved October 11st, 2010 from http://www.agrodesierto.
com/moringa.html.
AOAC, 1984. Official methods of analysis. 14th edition, (Association
of official analytical chemists, Washington)
AOAC, 1990. Official methods of analysis. 15th edition, (Association
of official analytical chemists, Gaithersburg)
Benavides, J., 1994. La investigacin en rboles forrajeros. rboles y
arbustos forrajeros de Amrica Central, (CATIE, Turrialba)
Bennet, R., Mellon, F., Foild, N., Pratt, J., Dupont, M., Perkins, L. and
Kroon, P., 2003. Profiling glucosinolates and phenolics in
vegetative and reproductive tissues of the multi-purpose trees
Moringa oleifera L. (Horseradish tree) and Moringa stenopetala
L, Journal of Agricultural and Food Chemistry, 51, 3546553
Ben Salem, H. and Makkar, H., 2009. Defatted Moringa oleifera seed
meal as a feed additive for sheep, Animal Feed Science and
Technology, 150, 2733
Crdenas, J., Sandoval, C. and Solorio, F., 2003. Chemical composition
of grass and forage trees mixed silages, Tcnica Pecuaria de Mxico,
41, 283294
Castro Ramrez, A., 2002. Ganadera de leche: Enfoque empresarial,
(Universidad Estatal a Distancia, San Jos)
de Leeuw, P., Omure, A., Staal, S. and Thorpe, W., 1998. Dairy
production systems in the tropics: A review. In: L. Falver and C.
Chantalakhan (eds), Smallholders dairying in the tropics, (ILRI,
Nairobi), 1944
Dougherty, R., Shipe, W., Gudnason, V., Ledford, R., Peterson, R. and
Scarpellino, R., 1962. Physiological mechanisms involved in
transmitting flavors and odors to milk. I. Contribution of eructated
gases to milk flavor, Journal of Dairy Science, 4, 472476
Duke, J., 1983. Handbook of energy crops (Moringa oleifera), (Center
for new crop and plant products, Purdue University, Indiana)
Fales, S., Gustine, D., Bosworth, S. and Hoover, R., 1987. Concen-
trations of glucosinolates and S-methylcysteine sulfoxide in
ensiled rape (Brassica napus L.), Journal of Dairy Science, 70,
24022405
FAO, 2010. Ganadera bovina en Amrica Latina: Escenario 2008
2009 y tendencias del sector, (Santiago de Chile)
Fernndez-Baca, S., 1992. Perspectivas de la produccin de leche y
carne en el trpico americano. In: S. Fernndez-Baca (ed),
Avances en la produccin de leche y carne en el trpico
Americano, (FAO, Santiago de Chile)
Trop Anim Health Prod (2011) 43:10391047 1045
Ferreira, P., Farias, D., Oliveira, J. and Carvalho, A., 2008. Moringa
oleifera: Bioactive compounds and nutritional potential, Revista
de Nutrio, 21, 431437
Fujihara, T., Kakengi, A., Shem, M. and Sarwatt, S., 2005. Can
Moringa oleifera be used as a protein supplement for ruminants?,
Asian-Australasian Journal of Animal Science, 18, 42
Holmann, F., Rivas, L., Carulla, C., Rivera, B., Giraldo, L., Guzmn,
S., Martnez, M., Medina, A. and Farrow, A., 2003. Evolution of
milk production systems in tropical Latin America and its
interrelationship with the markets: an analysis of the Colombian
case, Livestock Research for Rural Development 15, Retrieved
December 7, 2010, from http://www.lrrd.org/lrrd15/9/holm159.
htm
Kokkonen T, Tesfa AT, Tuori M, Yrjnen S, Syrjl-Qvist L., 2002.
Effect of concentrate crude protein level on grass silage intake,
milk yield and nutrient utilisation by dairy cows in early
lactation. Archiv fr Tierernhrung, 56, 21327.
Lindgren, E., 1979. Vallfodrets nringsvrde bestmt in vivo och med
olika laboratoriemetoder. Report 45, (The Department of Animal
Nutrition and Management, The Swedish University of Agricultural
Sciences, Uppsala)
Makkar, H. and Becker, K., 1996. Nutritional value and anti-
nutritional components of whole and ethanol extracted Moringa
oleifera leaves, Animal Feed Science and Technology, 63, 211
228
Makkar, H. and Becker, K., 1997. Nutrients and antiquality factors in
different morphological parts of the Moringa oleifera tree,
Journal of Agricultural Science, 128, 311332
Manh, L., Dung, N. and Ngoi, T., 2005. Introduction and evaluation of
Moringa oleifera for biomass production and as feed for goats in
the Mekong Delta, Livestock Research for Rural Development,
17, Retrieved December 7, 2010, from http://www.lrrd.org/
lrrd17/9/manh17104.htm
Mauricio, R., Sousa, L., Moreira, G., Reis, G. and Gonalves, L.,
2008. Silvopastoral systems as a sustainable alternative to animal
production in the tropics. In: O. Casteln, A.R. Bernus and F.
Mould (eds), Opportunities and challenges for smallholder
ruminant systems in latin America. Resources management, food
safety, quality and market access, (Universidad Autnoma del
Estado de Mxico, Toluca) 187200
Mendieta-Araica, B., Sporndly, E., Reyes-Snchez, N., Norell, L. and
Sporndly, R., 2009. Silage quality when Moringa oleifera is
ensiled in mixtures with Elephant grass, sugar cane and molasses,
Grass and Forage Science, 64, 364373
Mendieta-Araica, B., Sprndly, R., Reyes-Snchez, N. and Sprndly,
E., 2010. Moringa (Moringa oleifera) leaf meal as a source of
protein in locally produced concentrates for dairy cows fed low
protein diets in tropical areas, Livestock Science, doi:10.1016/j.
livsci.2010.09.021
Nash, J., 1985. Crop conservation and storage in cool, temperate
climates. (Pergamon Press, Oxford)
NRC, 1988. Nutrient Requirements of Dairy Cattle. Sixth revised
edition. National Research Council. National Academy Press.
Washington DC
NRC, 2001. Nutrient Requirements of Dairy Cattle. Seventh revised
edition. National Research Council. National Academy Press.
Washington DC
Oldham, J., 1984. Protein-Energy interrelationships in dairy cows,
Journal of Dairy Science, 67, 10901114
Panciera, M., Kunkle, W. and Fransen, S., 2003. Minor silage crops.
In: D. Buxton, R. Muck and J, Harrison (eds), Silage science and
technology, (American society of agronomy Inc., Crop science
society of America Inc., Soil science society of America Inc.,
Madison), 781824
Panditharatne, S., Allen, V., Fontenot, J. and Jayasuriya, M., 1986.
Ensiling characteristics of tropical grasses as influenced by stage
of growth, additives and chopping length, Journal of Animal
Science, 63, 197
Paterson, R., Karanja, G., Roothaert, R., Nyaata, O. and Kariuki, W.,
1998. A review of tree fodder production and utilization within
smallholder agroforestry systems in Kenya, Agroforestry Systems,
41, 191199
Patterson, H. and Lucas, H., 1962. Change-over design. Technical
Bulletin 147, (United States Department of Agriculture, North
Carolina)
Pezo, D., 1991. La calidad nutritiva de los forrajes. Produccin y
utilizacin de forrajes en el trpico, Compendio. Serie de
materiales de enseanza 15, (CATIE, Turrialba)
Phiri, M., Ngongoni, N., Maasdorp, B., Titterton, M., Mupangwa, J.,
Sebata, A. 2007. Ensiling characteristics and feeding value of
silage made from browse tree legume-maize mixtures. Tropical
and Subtropical Agroecosystems. 7, 149156.
Reyes-Snchez, N., 2006. Moringa oleifera and Cratylia argentea:
Potential fodder species for ruminants in Nicaragua, Doctoral
thesis No. 2001:1. (SLU Acta Universitatis Agriculturae Sueciae,
Uppsala)
Reyes-Snchez, N., Ledin, S. and Ledin, I., 2006a. Biomass production
and chemical composition of Moringa oleifera under different
management regimes in Nicaragua, Agroforestry Systems, 66,
231242
Reyes-Snchez, N., Sporndly, E., and Ledin, I., 2006b. Effect of
feeding different levels of foliage of Moringa oleifera to creole
dairy cows on intake, digestibility, milk production and compo-
sition, Livestock Science, 1001, 2431
Reyes-Snchez, N., Mendieta-Araica, B., Faria, T. and Mena, M.,
2008. Gua de suplementacin alimenticia estratgica para
bovinos en poca seca, Serie Guas Tcnicas 12 (Universidad
Nacional Agraria, Managua)
Reyes-Snchez, N., Mendieta-Araica, B., Faria, T., Mena, M.,
Cardona, J. and Pezo, D., 2009. Elaboracin y utilizacin de
ensilajes en la alimentacin del ganado bovino, Manual Tcnico
91 (CATIE, Managua)
Ruiz, M.E., 2005. Overcoming feed limitations in dual-purpose cattle
system in Latin America. In: A.A. Ayantunde, S. Fernndez-
Rivera and G. McCrabb (eds), Coping with feed scarcity in
smallholder livestock systems in developing countries, (ILRI,
Nairobi), 183208
SAS, 2004. User guide, version 9.1.2, (Statistical analysis system
institute Inc., Cary)
Sarwatt, S., Milang'ha, M., Lekule, F. and Madalla, N., 2004. Moringa
oleifera and cottonseed cake as supplements for smallholder
dairy cows fed Napier grass, Livestock Research for Rural
Development, 16, 6, Retrieved December 7, 2010, from http://
www.lrrd.org/lrrd16/6/sarw16038.htm
Schingoethe, D., 1996. Dietary influence on protein level in milk and
milk yield in dairy cows, Animal Feed Science and Technology,
60, 181190
Ser, C., Steinfeld, H. and Groenewold., J., 1995. World livestock
production systemsCurrent status, issues and trends (FAO,
Rome)
Shipe, W., Ledford, A., Peterson, R., Scanlan, R., Geerken, H.,
Dougherty, R. and Morgan, M., 1962. Physiological mechanisms
involved in transmitting flavors and odors to milk. II. Transmis-
sion of some flavor components of silage, Journal of Dairy
Science, 4, 477480
Shipe, W., Bassette, E., Deane, D., Dunkley, W., Hammond, E.,
Harper, W., Kleyn, D., Morgan, M., Nelson, J. and Scanlan, R.,
1978. Off flavors of milk: Nomenclature, standards and bibliog-
raphy, Journal of Dairy Science, 61, 855869
Steinfeld, H., Gerber, P., Wassenaar, T., Castel, V., Rosales, M. and de
Haan, C., 2006. Livestock's long shadow: environmental issues
and options, (FAO, Rome)
1046 Trop Anim Health Prod (2011) 43:10391047
Sutton, J., 1989. Altering milk composition by feeding, Journal of
Dairy Science, 72, 28012814
Szott, L., Ibrahim, M. and Beer, J., 2000. The hamburger connection
hangover: Cattle, pasture land degradation and alternative use in
Central America, (CATIE, Turrialba)
Tjandraatmadja, M., Norton, B. and Mac Rae, I., 1994. Ensilage
characteristics of three tropical grasses as influenced by stage of
growth and addition of molasses, World Journal of Microbiology
and Biotechnology, 10, 7481
Undersander, D., Mertens, D. and Theix, N., 1993. Forage
analyses procedures, (National forage testing association,
Omaha)
Valenciaga, D., Chongo, B. and La O, O., 2001. Characterization of
clon Pennisetum CUBA CT-115. Chemical composition and dry
matter ruminal degradability, Cuban Journal of Agricultural
Science, 35, 349354
Valenciaga, D., Chongo, B., Herrera, R., Torres, V., Oramas, A. and
Herrera, M., 2009. Effect of regrowth age on in vitro dry matter
digestibility of Pennisetum purpureum cv. CUBA CT-115, Cuban
Journal of Agricultural Science, 43, 7982
Van Soest, P., Robertson, J., and Lewis, B., 1991. Methods for dietary
Fiber, neutral-detergent Fiber and non-starch polysaccharides in
relation to animal nutrition, Journal of Dairy Science, 4, 35833597
Vlez, M., 1997. Produccin de Ganado lechero en el Trpico,
(Escuela Agrcola Panamericana, El Zamorano, Tegucigalpa)
Witting de Penna, E., 1981. Evaluacin sensorial, una metdica que
mide calidad. II. Evaluacin mediante el test de valoracin con
escala de Karlsruhe, Alimentos, 6, 2531
Trop Anim Health Prod (2011) 43:10391047 1047