Content uploaded by Renkang Peng
Author content
All content in this area was uploaded by Renkang Peng on Nov 03, 2014
Content may be subject to copyright.
The weaver ant,
Oecophylla smaragdina
(Hymenoptera: Formicidae), an effective
biological control agent of the red-banded thrips,
Selenothrips rubrocinctus
(Thysanoptera: Thripidae) in mango crops in the Northern Territory of Australia
(Keywords: weaver an t, r ed-ba nded thrips, biological con trol , man go, Oecophylla smaragdina, Seleno thr ips rubrocinctus )
R. K. PENG* and K. CHRISTIAN
School of Science, Charles Darwin University, Darwin NT 0 909, Australia
Abstract. Weaver ants, Oecophylla smaragdina (Fabricius), have
been successfully used to control the main insect pests of cashew
plantations in northern Australia and Papua New Guinea. The red-
banded thrips, Selenothrips rubrocinctus (Giard), is an economically
important insect pest of mango, Mangifera indica L., orchards in the
Northern Territory. This w ork was undertaken to evaluate whether
weaver ants, which are abundant in mango orchards, have the potential
to control the red-banded thrips. Field surveys, field experiments and
laboratory trials were carried out in four mango orchards in the Darwin
area over four years. In field surveys, the number of shoots damaged by
the thrip s was significantly lower on trees with abundant weaver ants
(2.8%) than with fewer ants (21.1%), or without the ants (30.3%). Trees
with abundant weaver ants also produced the highest numbers of flower
panicles. Leaf examinations revealed that newly mature leaves on trees
with abundant weaver ants had significantly fewer thrips than on trees
with fewer or no ants. Field experiments showed that weaver ants were
as effective as chemical insecticides in limiting fruit damage by thrips. In
laboratory trials, seedlings without weaver ants were heavily damaged,
and lost all their leaves within six weeks, while seed lings with weaver
ants grew well and lost no leaves. This work suggests that the weaver
ant is an effective biological control agent of the red-banded thrips, and
the use of weaver ants in mango orchards is discussed.
1. Introduction
The red-banded thrips, Selenothrips rubrocinctus (Giard)
(Thysanoptera: Thripidae), is one of the world’s major insect
pests, damaging a range of tropical and sub-tropical tree crops,
such as mango, cashew, avocado, cacao, guava, mangosteen,
rambutan, tung-oil tree and many kinds of ornamental trees
(Callan, 1975, Bennett and Baranowski, 1982, Igboekwe, 1985,
Dennill, 1992, Patel et al., 1997). It is an economically important
pest in orchards of mango, Mangifera indica L., in the Northern
Territory of Australia (Poffley, 1996, Young and Poffley, 1997,
Young and Chin, 1998). Red-banded thrips are normally abundant
between April and July (Young and Chin, 1998), when they cause
serious damage to newly mature leaves, and they also damage
new leaf flush and fruits, resulting in fallen leaves, denuded trees
and inferior fruits. More importantly, leaf damage in the pre-
flowering flush can lead to a significant reduction of flower panicles
produced by trees (Malcolm Green, Anna Couttie, Hannah
Couttie, Les Brigden and Diane Lucas, 2001, pers. comm.).
Red-banded thrips usually occur in mango orchards together
with other insect pests such as leafhoppers, fruit-spotting bugs,
leaf beetles and caterpillars. To control red-banded thrips and
other insect pests, conventional mango growers in the Northern
Territory rely on chemical insecticides such as Trichlorfon and
Dimethoate, using 6 – 8 sprays between March and August. The
pest populations can be controlled, but the heavy use of
insecticides has resulted in increased costs, the reduction of
natural predators and parasitoids of the insect pests, increased
insect pest resistance to insecticides and environmental pollution
(Ian Baker, pers. comm.). Some organic mango growers use
commercial predators (e.g. the lacewing, Mallada signata
(Schneider), (Neuroptera: Chrysopidae) to control the thrips.
Adequate control can be achieved if the predators are released
in sufficient numbers at the right time, but this operation is very
expensive due to the costs of the predators ($Aus550/ha for two
releases) and the required monitoring programme. Other organic
growers take no action against the thrips, and their trees are
damaged, resulting in reduction of yield and fruit quality (Malcolm
Green, Anna Couttie, Hannah Couttie, and George Sohn, 2001,
pers. comm.). Thus, taking no action against thrips is not
economically viable.
Weaver ants, Oecophylla smaragdina (Fabricius) (Hyme-
noptera: Formicidae), are known to control over 40 species of
insect pests on many tropical tree crops (Way and Khoo, 1992,
Peng et al., 1995, 2000a). This ant lives in leaf nests in the
canopy of many tropical trees, and it feeds on sugar-rich
materials and a range of insects by patrolling various parts of
trees. Barzman et al. (1996), Van Mele and Cuc (2000) and Van
Mele et al. (2002) suggested that weaver ants could be used in
citrus orchards in Vietnam. Since 1998, weaver ant colonies
have been successfully used to control the main insect pests in
cashew orchards in the Northern Territory and Papua New
Guinea (Peng et al., 1999, 2000b, Peng, 2001). This species of
ant occurs abundantly in mango orchards in the Northern
Territory, and therefore, they may have the potential to reduce
red-banded thrips populations.
2. Materials and Methods
Four mango orchards in the Darwin (128 40’S 130 8 81’E)
area of the Northern Territory were used in this study in 1996,
1997, 2001 and 2002. Orchards A, B and C are 22 km, 29 km
and 23 km respectively south-east of Darwin, and orchard D is
*To whom correspondence should be addressed. e-mail: renkang.peng@cdu.edu.au
INTERNATIONAL JOURNAL OF PEST MANAGEMENT, APRIL–JUNE 2004, 50(2) 107–114
International Journal of Pest Management
ISSN 0967-0874 pr int/I SSN 1366-5863 online
#
2004 T aylor & Fran cis Ltd
http://ww w.tand f.co .uk/ jour nals
DOI: 10.1080/096708704100 016581 25
38 km south-south-east of Darwin. The trees were between four
and seven years old, and were of the Kensington Pride variety.
Field surveys, field experiments and laboratory trials were used
in this study.
2.1. Field survey (assessment of thrips numbers)
Two field surveys (one survey in March and the other in
May) were performed in orchard A in 1996. In March, when
the red-banded thrips populations were increasing, every tree
in a mango block of 0.5 ha was sampled. A total of eight
newly mature leaves (most recently hardened flush) were
picked from four sides of a tree (two on each side) using a
picking pole, and each leaf was put in a small plastic bag. The
bagged samples were immediately taken to the laboratory and
the red-banded thrips (nymphs and adults) on each leaf and in
the plastic bag were counted under a binocular microscope.
Weaver ant abundance on each tree was assessed at the
same time. The number of ant trails on the main branches of a
tree was counted while tapping the tree trunk with a stick, and
the total number of main branches on the tree was also
recorded. The percentage of the main branches with ant trails
was calculated for each tree. Weaver ants on a tree were
treated as ‘Abundant’, if 5 50% of the main branches had ant
trails, or as ‘Fewer ants’ if 5 50% of the main branches had
ant trails. Some trees, which were heavily foraged by the
ground nesting meat ant, Iridomyrmex sanguineus (Forel)
(Hymenoptera: Formicidae), were recorded as ‘trees with meat
ants’.
In May, when the thrips populations were large, every tree in
a block of 0.7 ha (including the 0.5 ha used in March) was
sampled. Five newly mature leaves per tree were picked (one on
each side and one at the top), and the same procedures that
were carried out in March were performed. Thrips numbers
among four categories (trees with abundant weaver ants, with
fewer weaver ants, with meat ants and without ants) were
analysed by a Kruskal-Wallis one-way ANOVA by ranks (non-
parametric) (Siegel, 1956) using SYSTAT statistical software
(Wilkinson, 1990).
2.2. Field survey (assessment of damaged shoots)
Red-banded thrips are abundant between April and July in
the Northern Territory (Young and Chin, 1998), and field
surveys were performed between May and July. A total of
eight field surveys were conducted: five in orchard A in 1996,
1997, 2001 and 2002, two in orchard B in 2002 and one in
orchard D in 2001. In each survey, every tree in an orchard was
inspected. For each tree, all the shoots in the outer-middle area
of the tree were examined, which is where the thrips are most
abundant (Young and Chin, 1998). The thrips prefer to feed on
the tissue next to the midrib on the undersurface of newly
mature leaves. The sign of fresh damage is silvering patches
on leaves with numerous small, shiny black spots of excreta,
and then the silvering develops a pale-yellow to brown
discolouration, speckled darkly with dried droppings. In the
assessment of shoot damage, two variables were used: the
total number of newly mature leaves on a shoot, and the
number of damaged leaves. A leaf was treated as ‘damaged’ if
more than 30% of the whole leaf area had signs of fresh
damage, otherwise the leaf was classified as ‘not damaged’.
Based on our field observations, when more than 30% of a leaf
area was damaged the leaf gradually turned brown, dried and
dropped from the tree. This did not occur when less than 30%
of a leaf area was infested by thrips. If the damaged leaves
represented more than 30% of the newly mature leaves on a
shoot, the shoot was recorded as ‘damaged’, otherwise it was
considered ‘not damaged’. According to our observations,
shoots often bear more than 10 newly mature leaves. If a
shoot had more than 30% of the leaves ‘damaged’, the shoot
gradually became denuded; otherwise the shoot did not
become denuded. Percentage of damaged shoots per tree
was calculated as follows: total number of damaged shoots/
total number of shoots in the outer-middle area of the
tree 6 100. The abundance of weaver ants and other species
of ants on each tree was assessed at the same time as the
assessment of leaf shoots, and the criteria used for the
assessment are given above. The damage by the thrips was
compared for groups of trees (those with abundant weaver
ants, fewer weaver ants, other species of ants, and no ants)
using the Kruskal – Wallis one-way ANOVA by ranks (Siegel,
1956).
2.3. Field experiment
Field experiments were done in orchards A and C in 2001
and 2002. In orchard A, during the mango pre-flowering flush in
May 2001, a 1.4 ha mango block was equally divided into three
replicates, and each replicate had two treatments. One
treatment had weaver ants present, and the other had no
weaver ants present but chemical insecticides were used. In
the weaver ant treatment, the ants attend mealybugs for honey
dew, and mealybug populations were high and caused damage
on fruits. To reduce this damage, one of two ‘soft’ chemicals
(Petroleum spray oil (D.C. Tron Plus) at 1.5% v/v or Potassium
Soap at 1% v/v), both of which have little effect on weaver ants,
was used when 5 – 10% of fruits were infested by mealybugs.
In the treatment with chemical insecticides, Lepidex 500 (500 g/
L Trichlorfon) at 0.1% v/v and/or Dimethoate (400 g/L Dimetho-
ate) at 0.1% v/v were used when insect pests caused a
damage of 5 – 10% of foliar or floral shoots or developing fruits
(acceptable levels by mango growers). Therefore, this block
had two treatments, and each treatment had three replicates,
each of which had between 32 and 43 mango trees of similar
size and age. Weaver ants were transplanted into each of the
three ‘with weaver ant’ replicates in May 2001. Monitoring of the
main insect pests and the ant abundance in each replicate
started early June 2001 and was done once a week during the
pre-flowering flush and flowering and fruiting flush and once a
month at other times. During the mango harvest, every fruit was
assessed on site, based on the Mango Quality Standards
produced by the Queensland Mango Sub-committee and
H.R.D.C. for the Mango Growers of Australia, which is the
standard assessment used by mango packing sheds in the
Northern Territory. If a fruit had an area greater than 3 cm
2
of
thrips damage (silvery patches with numerous small, shiny
black spots of excreta) on its skin, the fruit was treated as
‘damaged’.
In orchard C, 38 mango trees occupying an area of 0.4 ha
were used. Based on the distribution of existing weaver ant
108 R. K. Peng and K. Christian
colonies, the orchard was divided into three treatments. The
first treatment included 12 trees, which had weaver ants
present and were treated with ‘soft’ insecticides (D.C. Tron
Plus at 1.5% v/v or Potassium Soap at 1% v/v was used when
5 – 10% of fruits were infested by mealybugs). The second
treatment included 13 trees, which had weaver ants present but
were not treated with chemicals. The third treatment included
13 trees, which had no weaver ants present and no soft
chemicals or chemical insecticides were used. The procedures
of monitoring and fruit quality assessment were the same as in
orchard A.
For orchard A, a two-way ANOVA was used to examine the
effect of two treatments (trees with weaver ants plus ‘soft’
chemicals and trees with chemical insecticides) in the three
replicates. An arcsine (square root) transformation was applied
to the percentage damage of fruit to achieve normal distributions
before the ANOVA was conducted. In orchard C, percentage of
fruit damage was analysed by the Kruskal-Wallis one-way
ANOVA by ranks (Siegel, 1956).
2.4. Laboratory trial
Two laboratory trials were performed at the Charles Darwin
University campus in 2001 and 2002 to determine the effects of
weaver ants on red-banded thrips. In 2001, cashew seedlings
were used to rear the thrips because (1) both cashew and mango
are red-banded thrips hosts, and (2) at the time of the trial,
mango seedlings were not available. Each flower pot had a
cashew seedling grown from seed, and 10 pots with seedlings
were prepared. When seedlings were 4 weeks old and had about
6 – 7 leaves, each seedling was inoculated with 26 – 60 nymphs
of red-banded thrips, depending on the number of leaves on the
seedling. Red-banded thrips were collected from mango trees in
orchard A. To transplant the thrips, a piece of leaf with a known
number of thrips larvae (which had been counted under a
binocular microscope) was attached to a leaf on a seedling. After
the inoculation, the ten seedlings were kept in an entomological
laboratory at 25 – 288C for 3 days to allow the thrips to move to
cashew leaves.
On the fourth day, the 10 pots were taken to a well-
established weaver ant colony. Each pot was put in a saucer
filled with water. Five of these pots were put on one side of the
main weaver ant trail, and the weaver ants were kept away
from the cashew seedlings by the water in the saucers. Another
five seedlings were put on the other side of the ant trail, and
these saucers were filled with a combination of water and
gravel, so the ants were able to walk on the gravel to reach the
seedlings. Because the seedlings were too small for the ants to
construct nests, two small weaver ant nests (which were taken
from the adjoining colony) were transferred onto each seedling
pot every week. The number of nymphs and adults of red-
banded thrips on each seedling and the growth of the cashew
seedlings were recorded prior to introducing new ant nests
each week.
In 2002, mango seedlings were used in the laboratory trial.
These seedlings were affected by red-banded thrips by the
time they had 6 – 8 leaves. The number of nymphs and adults
of red-banded thrips on each of the seedlings was counted at
the beginning of the trial. The procedures of this trial were the
same as the 2001 trial. Unfortunately, this experiment was
stopped in week 5 because irrigation was accidentally started,
which affected the thrips numbers on the seedlings and the
foraging behaviour of the ants. As data for both trials did not
meet the requirement of parametric t-test assumptions, the
data were analysed by the Mann – Whitney U-test (Siegel,
1956) using non-parametric statistics software (Wilkinson,
1990).
3. Results
3.1. Field survey (assessment of thrips numbers)
In orchard A in March 1996, the average number of thrips
found on a new mature leaf was significantly lower in trees with
abundant weaver ants (0.3) than in trees with fewer weaver ants
(2.9), or in trees with meat ants (1.4), or in trees without ants (3.5)
(table 1). Similar results were observed in May when the red-
banded thrips populations were high (table 1).
3.2. Field survey (assessment of damaged shoots)
In orchard A, data from May 1996 demonstrated that 3.8% of
shoots in trees with abundant weaver ants were damaged by
red-banded thrips, which was much lower than in trees with meat
ants (27.1%) or in trees without ants (64.7%) (table 2a). Data
from surveys done in July 1997 and in June 2001 revealed that
trees with abundant weaver ants were much less damaged by
the t hrips than trees with fewer weaver ants or without weaver
ants (table 2). This characterization holds true for the survey
done in June and July 2002 (table 3a). Also, in the July 2002
survey, trees with abundant weaver ants produced the most
flower panicles (table 3a).
Table 1. Populations of red-banded thrips in orchard A in 1996
15 March 1996 3 May 1996
Trees with Mean number of thrips/leaf + SD
1
N
2
Mean number of thrips/leaf + SD N
Abundant Oecophylla smaragdina 0.3 + 0.6 208 2.8 + 3.8 150
Fewer O. smaragdina 2.9 + 4.9 48 3.2 + 2.9 50
Iridomyrmex sanguineus 1.4 + 2.2 168 17.0 + 18.8 65
No ants 3.5 + 2.7 48 12.6 + 8.3 145
Kruskal – Wallis test H = 19.4, df = 3, p 5 0.001 H = 28.3, df = 3, p 5 0.001
1
SD refers to the standard deviation of the mean.
2
N refers to the number of newly mature leaves examined.
109Effects of
Oecophylla smaragdina
on
Selenothrips rubrocinctus
Similarly in orchard B, trees with abundant weaver ants were
also significantly less damaged by red-banded thrips than trees
with fewer weaver ants or without weaver ants (table 3b). In
addition, trees with abundant weaver ants also produced the
highest number of flower panicles (table 3b).
In orchard D, trees with abundant weaver ants were
damaged significantly less by red-banded thrips than trees with
other species of ants (table 2b).
3.3. Field experiment
In 2001, the thrips damaged a higher proportion of the fruit
from trees treated with chemical insecticides (3.9%) compared
with trees with weaver ants plus ‘soft’ chemicals (0.9%, table 4a).
There was no difference among replicates (p = 0.76) and no
significant interaction between treatments and replicates
(p = 0.75). In 2002, there was no difference in the proportion of
Table 2. Damage caused by red-banded thrips in (a) orchard A in 1996 and 1997 and (b) two orchards in June 2001
(a)
May 1996 July 1997
Trees with Mean % damaged shoots/tree + SD
1
N
2
Mean % damaged shoots/tree + SD N
Abundant Oecophylla smaragdina 3.8 + 1.3 32 1.3 + 2.8 24
Fewer O. smaragdina – 3.7 + 5.8 13
Iridomyrmex sanguineus 27.1 + 4.9 21 –
No ants 64.7 + 10.1 6 19.2 + 23.5 20
Kruskal – Wallis test H = 35.4, df = 2, p 5 0.001 H = 10.3, df = 2, p 5 0.01
(b)
Orchard A Orchard D
Trees with Mean % damaged shoots/tree + SD
1
N
2
Mean % damaged shoots/tree + SD N
Abundant O. smaragdina 5.2 + 8.0 22 2.6 + 2.6 12
Fewer O. smaragdina 28.4 + 24.3 10 –
Iridomyrmex sanguineus 6.2 + 8.5 5 17.5 + 13.1 4
Iridomyrmex sp – 35.2 + 29.9 18
No ants 41.2 + 32.6 8 –
Kruskal – Wallis test H = 18.3, df = 3, p 5 0.001 H = 17.2, df = 2, p 5 0.001
1
SD refers to standard deviation of the mean.
2
N refers to the number of trees examined.
Table 3. Damage caused by red-banded thrips in (a) orchard A and (b) orchard B in 2002
(a)
5 June 2002 3 July 2002
Trees with
Mean % damaged
shoots/tree + SD
1
N
2
Mean % damaged
shoots/tree + SD
Number of flower
panicles/tree + SD N
Abundant Oecophylla smaragdina 0.5 + 1.1 75 3.1 + 7.1 63.2 + 52.2 55
Fewer O. smaragdina 13.0 + 14.7 20 36.4 + 27.0 8.8 + 9.0 29
Iridomyrmex sanguineus 8.6 + 11.8 4 23.6 + 10.7 52.9 + 20.6 7
Iridomyrmex sp 19.4 + 18.3 5 48.0 + 37.4 22.8 + 15.1 5
No ants 33.2 + 23.6 22 56.1 + 25.2 4.4 + 6.3 18
Kruskal – Wallis test H = 79.5, df = 4, p 5 0.001 H = 71.0, df = 4, p 5 0.001 H = 68.6, df = 4, P 5 0.001
(b)
28 May 2002 10 July 2002
Trees with
Mean % damaged
shoots/tree + SD
1
N
2
Mean % damaged
shoots/tree + SD
Number of flower
panicles/tree + SD
N
Abundant O. smaragdina 1.7 + 3.2 46 6.6 + 9.9 91.7 + 67.5 46
Fewer O. smaragdina 13.6 + 15.6 7 18.6 + 25.4 28.7 + 17.4 7
Iridomyrmex sp 38.6 + 18.5 6 22.2 + 19.3 63.8 + 48.3 7
No ants 29.6 + 26.6 18 26.6 + 18.5 26.9 + 29.2 17
Kruskal – Wallis test H = 38.9, df = 3, p 5 0.001 H = 23.9, df = 3, P 5 0.001 H = 28.9, df = 3, P 5 0.001
1
SD refers to standard deviation of the mean.
2
N refers to the number of trees examined.
110 R. K. Peng and K. Christian
fruit damaged from trees treated with chemical insecticides
(0.4%) compared to fruits from trees with weaver ants plus ‘soft’
chemicals (0.2%, table 4a). However, in this year there was a
significant difference among replicates (p = 0.02), but the inter-
action between treatments and replicates was not significant
(p = 0.09).
In orchard C in both years, fruits were significantly less
damaged by the red-banded thrips in the treatments with weaver
ants only or with weaver ants plus ‘soft’ chemicals than in the
treatment without weaver ants or soft chemicals (table 4b).
3.4. Laboratory trials
In the 2001 trial with cashew seedlings, the number of
nymphs transplanted on seedlings at the beginning was similar
between two groups of seedlings (20 August, table 5). In the first
2 weeks of the trial, the majority of nymphs became pupae and
emerged as adults. The number of nymphs was similar between
the two groups of seedlings, but adult numbers were much
higher on the seedlings without weaver ants than with weaver
ants (27 August and 5 September, table 5). Starting from 10
September, nymphs started to hatch from eggs, and thrips
generations started to overlap from 28 September. During the
period between 10 September and 12 October, nymph and adult
numbers were significantly higher on seedlings without weaver
ants than with weaver ants (table 5). Starting from the third week
of this trial, the mean number of thrips per seedling was well
below 250 individuals on seedlings with weaver ants, while on
seedlings without weaver ants, the thrips numbers were always
well above 300.
In the first two weeks of the trial the number of healthy and
damaged leaves per seedling was not different between the two
groups of seedlings (27 August and 5 September, table 5).
Starting from week 3 (10 September), healthy leaves were
generated each week on seedlings with weaver ants. In
contrast, fewer healthy leaves were found on seedlings without
weaver ants, and more leaves were damaged each week (table
5). The damaged leaves on seedlings without weaver ants
started to fall from week 4 of the trial, and all the leaves on
these seedlings had fallen by 12 October (table 5). This
difference in the number of leaves fallen in the two groups was
highly significant (table 5).
In the 2002 trial with mango seedlings, the initial count of
thrips was not different between two groups of seedlings (10
May, table 6). Two weeks later and onwards, the mean number
of nymphs and adults per seedling was significantly smaller on
seedlings with weaver ants than without the ants (22 May – 6
June, table 6). The mean number of the thrips per seedling was
under 150 on seedlings with weaver ants, while the thrips on
seedlings without the ants exceeded 200, except for 29 May
(table 6). By week 4 (6 June), the mean number of healthy leaves
per seedling was significantly greater on seedlings with weaver
ants than without the ants (table 6). Seedlings with the ants had
fewer newly damaged leaves than those without the ants. From
the third week of the trial, the damaged leaves on seedlings
without the ants started to fall, and more leaves had fallen by the
fourth week (table 6). No leaves had fallen from the seedlings
with the ants.
4. Discussion
A range of natural enemies of the red-banded thrips has
been reported on various tree crops. These include a parasitoid
wasp, Goetheana parvipennis (Gahan) (Hymenoptera: Eulophi-
dae) on mango (Bennet and Baranowski, 1982), an anthocorid
bug, Orius thripoborus (Hesse) (Hemiptera: Anthocoridae) on
avocado (Dennill, 1992), and three species of Miridae, Terma-
tophylidea maculata (Usinger), T. pilosa (Reut. & Popp.) and T.
opaca (Carvalho) on cacao and cashew (Callan, 1975). In the
Northern Territory, Young and Chin (1998) mentioned that red-
banded thrips were attacked by spiders, lacewings, predatory
thrips and predatory bugs. A wasp, Shakespearia sp. (Hyme-
noptera: Encyrtidae) parasitises pupae of the thrips in the
Northern Territory (Lanni Zhang, 2003, pers. comm.). The
weaver ant has not been reported as a natural enemy of the
red-banded thrips.
Table 4. Red-banded thrips damage on mango fruits in different treatment in (a) orchard A and (b) orchard C
(a)
2001 2002
Treatment Replicates % fruit damage + SD
1
N
2
% fruit damage + SD N
Chemical insecticide 3 3.9 + 6.8 66 0.4 + 1.2 117
Oecophylla smaragdina and soft chemicals 3 0.9 + 2.7 45 0.2 + 0.7 80
ANOVA F = 10.8, df = 1, p = 0.001 F = 0.4, df = 1, p = 0.518
(b)
2001 2002
Treatment % fruit damage + SD
1
N
2
% fruit damage + SD N
Oecophylla smaragdina only 0.1 + 0.2 13 0.1 + 0.2 12
O. smaragdina and soft chemicals 0.1 + 0.2 11 0.0 + 0.1 12
No O. smaragdina and no soft chemicals 1.3 + 1.5 11 0.8 + 0.8 13
Kruskal – Wallis test H = 11.7, df = 2, p = 0.003 H = 10.1, df = 2, p = 0.006
1
SD refers to standard deviation of the mean.
2
N refers to the number of trees assessed.
111Effects of
Oecophylla smaragdina
on
Selenothrips rubrocinctus
Table 5. The effect of Oecophylla smaragdina on the development of red-banded thrips on cashew seedlings, at the Charles Darwin University campus, August – October, 2001
Category Measurement Seedlings with 20 Aug. 27 Aug 5 Sep. 10 Sep. 13 Sep. 20 Sep. 28 Sep. 5 Oct. 12 Oct.
Development Mean numbers of O. smaragdina 40 + 17 1 + 33+ 541+ 43 57 + 54 83 + 55 35 + 7 247 + 136 89 + 27
of thrips nymphs/seedling + SD
1
No ants 38 + 51+ 15+ 10 309 + 75 323 + 26 630 + 259 262 + 103 515 + 303 836 + 400
M-W U test
2
NS NS NS ** ** ** * NS *
Mean numbers of O. smaragdina 0 + 04+ 52+ 22+ 14+ 313+ 922+ 39+ 56+ 3
adults/seedling + SD No ants 0 + 021+ 10 8 + 67+ 39+ 3 168 + 42 272 + 29 46 + 14 15 + 19
M-W U test NS ** * NS **** *NS
Development Mean numbers of healthy O. smaragdina 6 + 27+ 212+ 313+ 315+ 317+ 318+ 320+ 321+ 2
of seedlings leaves/seedling + SD No ants 7 + 18+ 212+ 28+ 36+ 46+ 24+ 21+ 10.0+ 0.0
M-W U test NS NS NS * * ** ** ** *
Mean numbers of damaged O. smaragdina 000001+ 11+ 22+ 12+ 1
leaves/seedling + SD No ants 0 0 0 3 + 14+ 1 leaves
started falling
5 + 2 more
leaves fallen
5 + 2 more
leaves fallen
4 + 2 most
leaves fallen
All leaves
fallen
M-W U test NS NS NS ** ** * * – –
1
Nymphs included pre-pupae and pupae, and SD refers to standard deviation of the mean.
2
M-W U test refers to Mann-Whitney U-test; NS = not significant, * = significant at p 5 0.05 and ** = significant at p 5 0.01.
Table 6. The effect of Oecophylla smaragdina on the development of red-banded thrips on mango seedlings, at the Charles Darwin University campus, May – June 2002
Category Measurement Seedlings with 10 May 16 May 22 May 29 May 6 June
Development Mean numbers of O. smaragdina 273 + 156 140 + 104 1 + 10+ 02+ 3
of thrips nymphs/seedling + SD
1
No ants 340 + 377 398 + 338 82 + 34 9 + 17 280 + 313
M-W U test
2
NS NS ** * **
Mean numbers of O. smaragdina 1 + 19+ 14 1 + 10+ 00+ 1
adults/seedling + SD No ants 3 + 337+ 41 126 + 135 16 + 32 5 + 8
M-W U test NS ***NS **
Development Mean numbers of healthy O. smaragdina 6 + 36+ 35+ 45+ 37+ 5
of seedlings leaves/seedling + SD No ants 9 + 510+ 67+ 56+ 50+ 0
M-W U test NS NS NS NS **
Mean numbers of newly damaged O. smaragdina 3 + 13+ 11+ 10+ 00+ 0
leaves/seedling + SD No ants 2 + 23+ 25+ 34+ 2 leaves started falling 6 + 3 more leaves fallen
M-W U test NS NS ** ** **
1
Nymphs included pre-pupae and pupae, and SD refers to standard deviation of the mean.
2
M-W U test refers to Mann-Whitney U-test; NS = not significant, * = significant at p 5 0.05 and ** = significant at p 5 0.01.
112 R. K. Peng and K. Christian
Weaver ants are effective in limiting damage to mango leaf
shoots by the red-banded thrips. Data from the field surveys in
each of three orchards (tables 2 – 3) were consistent, and
demonstrated that trees with abundant weaver ants were
significantly less damaged than trees with fewer, or without,
weaver ants. Based on the data from eight field surveys in three
orchards over the period of four years, the mean damage in trees
with abundant weaver ants was 2.8% of the shoots, while 21.1%
and 30.3% of the shoots were damaged on trees with fewer
weaver ants or without the ants respectively. Damage levels of
5 5% of shoots damaged by red-banded thrips are acceptable to
farmers (Les Brigden and Malcolm Green, 2003, pers. comm.).
Weaver ants are effective in reducing thrips numbers.
Whether the red-banded thrips populations were high or low in
orchards, the number of the thrips per leaf was lowest on trees
with abundant weaver ants (table 1). Igboekwe (1985) studied
cashew seedlings infested by red-banded thrips and suggested
that when the thrips populations built up to 240 thrips per
seedling, chemical treatment was necessary, and 240 thrips per
seedling appeared to be an economic injury level. In our
laboratory trials, cashew and mango seedlings with weaver ants
were healthy and no leaves dropped (tables 5 – 6); the mean
number of the thrips per seedling, including nymph s and adults,
was well below 250 (tables 5 – 6). This suggests that weaver ants
can limit red-banded thrips populations under the economic
injury level for mango and cashew seedlings as that was
suggested by Igboekwe (1985).
Weaver ants are effective in protecting mango fruits from red-
banded thrips. The proportion of fruits damaged by the thrips in
orchard C in each of the two years was much lower on trees with
weaver ants than without the ants (table 4). In orchard A in 2001,
the proportion of fruits damaged by the red-banded thrips was
higher in the treatment with chemical insecticides than in the
treatment with weaver ants plus ‘soft’ chemicals (table 4a). In
2002, there was no significant difference of the fruit damage
between the two treatments, suggesting that weaver ants are at
least as effective as the chemical insecticides used in these
trials.
There are two species of weaver ants in the world:
Oecophylla smaragdina (Fab.) distributed widely in south-east
Asian countries, south Pacific islands and northern Australia; and
O. longinoda (Latreille) distributed in tropical Africa (Cole and
Johns, 1948). These species are so similar in their life-histories
and ecology that they can be treated as one (Way, 1954, Van der
Plank, 1960, Greenslade, 1971a, Holldobler and Wilson, 1983,
Peng et al., 1998a,b, Way and Khoo, 1992). Mangos grown in
the areas of tropical Asia, Australasia and Africa suffer from red-
banded thrips damage (Hill, 1975, Pena, et al., 2002). Most
nations in these areas are developing countries, where
insecticides and spray equipment are usually expensive and
labour costs are low. Thus, the use of the ants to control red-
banded thrips can be both successful and cost-effective.
Compared to other natural enemies of red-banded thrips,
weaver ants have the following advantages:
(1) Under natural conditions, weaver ants can significantly
reduce red-banded thrips damage in mango orchards
(tables 1 – 3). Weaver ants can also reduce red-banded
thrips numbers under the economic injury level on mango
and cashew seedlings;
(2) Mango trees are ideal hosts for weaver ants, and the ants
are often abundant in unsprayed mango orchards (Majer
and Camer-Pesci, 1991). In our field surveys, an average
of 47% of mango trees (tables 1 – 3) had abundant
weaver ants. Therefore, it is easy to locate and maintain
the ants in mango orchards;
(3) The ants can live on mango trees all year, constantly
controlling red-banded thrips populations;
(4) For mango orchards without weaver ants, transplantation
of the ants into the orchards is possible. A transplanted
ant colony with queen ants can last three or more years in
orchards (Peng et al., 2000b). By comparison, if other
natural enemies such as parasitoid wasps and lacewings
are used, a repeated release of the wasps and lacewings
at least twice a year is necessary; and
(5) Apart from controlling the red-banded thrips, weaver ants
can also control several other main insect pests in mango
orchards, such as leafhoppers, leaf-rollers, seed weevils,
fruit-spotting bugs, flower caterpillars and fruit-flies (Peng
and Christian, unpublished data), but the other natural
enemies cannot control this range of pest species.
However, there are several arguments against using the ants
in mango orchards. These include the mutual relationship
between the ants and mealybugs, the fact that fruit quality can
be affected by drops of formic acid secreted by the ants, the
aggressive behaviour of the ants and the ant population stability
in orchards.
Weaver ants have a close association with some homopter-
ans such as scales, mealybugs and some species of aphids
(Way, 1963), and they encourage mealybug populations in
mango orchards (Peng and Christian, unpublished data).
According to the results of our current research, the mean fruit
damage by all insect pests, including caterpillars, thrips, fruit-
spotting bugs, seed weevils, fruit-flies and mealybugs, was 10%
in trees protected by weaver ants plus ‘soft’ chemicals, in which
mealybug damage accounted for 3%. In trees protected by
chemical insecticides, 13% of fruits were damaged on average,
in which mealybug damage accounted for 0.4% (Peng and
Christian, unpublished data). This suggests that mealybug
damage on mango fruits resulting from the use of weaver ants
is balanced by less damage by other insect pests.
The deposition of weaver ant formic acid can cause black
spots on fruit skin. A fruit is downgraded if there are 4 10 black
spots on it. In our field experiments, 3% of fruits were
downgraded due to ant marks (Peng and Christian, unpublished
data). If fruits damaged by all insect pests, including ant marks,
are considered, 13% of fruits were damaged in trees protected
by weaver ants plus ‘soft’ chemicals, while 13% of fruits were
also damaged in trees protected by chemical insecticides. This
suggests that the damage on mango fruits due to ant marks is
compensated by the beneficial effects of the ants.
Aggressive behaviour is well known in weaver ants, and this
may disturb people during fruit harvest. Spraying with contact-
killing insecticides like Pyrethrum in mango orchards reduces the
abundance of weaver ants, but it also reduces ant ability to
control pests. During the period of mango harvest in 2001 and
2002, six farmers who are involved in this research did not
consider ant aggressiveness a significant problem. To avoid
disturbance from the ants, some farmers put fruits directly into a
113Effects of
Oecophylla smaragdina
on
Selenothrips rubrocinctus
bucket filled with water and mango wash solution for 1 min. This
also serves to remove sap from the fruits. Farmers in Vietnam
and Gabon sometimes put ash powder on the tree trunk or on
main branches when they harvest fruits or temporarily remove
ant nests (Van Mele and Cuc, 2003). Based on our observations
that weaver ants either went back to their nests or stayed under
twigs and leaves when it was raining, a field experiment using
seven-year-old mango trees showed that the activity of weaver
ants was reduced by 88% for the first 20 min after spraying with
water and by 61% for further 30 min (Peng and Christian,
unpublished data). Mango fruits and tree leaves dried in 20 min,
thus the added water was too little to cause fruit lenticel or post-
harvest diseases (Anna Couttie, Les Brigden, Malcolm Green
and Lloyd Pierce, 2003, pers. comm.).
Under natural conditions, weaver ant populations are not
distributed evenly across trees and months of the year in mango
orchards. This distribution affects the control efficiency of red-
banded thrips (tables 2 and 3). Population fluxes are related to the
fact that weaver ants have strong territorial behaviour. Fierce
boundary fights between weaver ant colonies (Peng et al., 1999)
and between weaver ants and other ant species (Greenslade,
1971b, Majer and Camer-Pesci, 1991, Way and Khoo, 1992,
Peng, personal observations) are well documented. To maintain
weaver ant populations at high levels to control red-banded thrips
in mango orchards, it is advisable to isolate weaver ant colonies
from each other, to use ant baits to control other competitive ant
species, and to transplant new weaver ant colonies as needed
(Way and Khoo, 1992, Peng et al., 1999, 2000b).
Acknowledgements
The study was supported by the Australian Centre for
International Agricultural Research and the Rural Industries
Research and Development Corporation. The authors are
grateful to Mr Les Brigden, Mr Malcolm Green, Ms Dallas Johns
and Mrs Diane Lucas for providing us with the study sites (A, B,
C and D respectively) and the necessary farm facilities. We
thank Mr Matthew Shortus, Ms Lanni Zhang and Mr Les Brigden
for their technical assistance.
References
BARZMAN, M. S., MILLS, N. J. and NGUYEN, T. T. C., 1996. Traditional
knowledge and rational for weaver ant husbandry in the Mekong delta of
Vietnam. Agriculture and Human Values, 13, 2–9.
BENNETT, F. D. and BARANOWSKI, R. M., 1982. First record of the thrips
parasite Goetheana parvipennis (Gahan) (Eulophidae: Hymenoptera) from
the Bahamas. Florida Entomologist, 65, 185.
CALLAN, E. M., 1975. Miridae of the genus Termatophylidea (Hemiptera) as
predators of cacao thrips. Entomophaga, 20, 389 – 391.
COLE, A. C. and JOHNS, J. W., 1948. A study of the weaver ant, Oecophylla
smaragdina (Fab.). American Midland Naturalist, 39, 641 – 651.
DENNILL, G. B., 1992. Orius thripoborus (Anthocoridae), a potential biocontrol
agent of Heliothrips haemorrhoidalis and Selenothrips rubrocinctus
(Thripidae) on Avocado fruits in the eastern Transvaal. Journal of the
Entomological Society of Southern Africa, 55, 255 – 258.
GREENSLADE, P. J. M., 1971a. Phenology of three ant species in the Solomon
Islands. Journal of Australian Entomological Society, 10, 241 – 252.
GREENSLADE, P. J. M., 1971b. Interspecific competition and frequency
changes among ants in the Solomom Islands coconut plantations. Journal
of Applied Ecology, 8, 323 – 353.
HILL, D., 1975. Agricultural insect pests of the tropics and their control.
(Cambridge: Cambridge University Press).
HOLLDOBLER, B. and WILSON, E. O., 1983. Queen control in colonies of
weaver ants (Hymenoptera: Formicidae). Annals of the Entomological
Society of America, 76, 235 – 238.
IGBOEKWE, A. D., 1985. Injury to young cashew plants, Anacardium occidentale
L., bythe red-banded thrips, Selenothrips rubrocinctus Giard (Thysanoptera:
Thripidae). Agriculture, Ecosystems and Environment, 13, 25 – 32.
MAJER, J. D. and CAMER-PESCI, P., 1991. Ant species in tropical Australian
tree crops and native ecosystems – Is there a mosaic? Biotropica, 23,
173 – 181.
PATEL, J. R., PATEL, M. B., RADADIA, G. C., SHAH, A. A. and PANDRA, H.
V., 1997. Insect pest management in mango nursery. Journal of Applied
Horticulture Navsari, 3, 125 – 128.
PENA, J., SHARP, J. and WYSOKI, M., 2002. Tropical fruit pests and
pollinators: Biology, economical importance, natural enemies and control.
(CABI Publishing), pp.448.
PENG, R. K., 2001. The control of cashew insect pests in cashew plantations
and small holder plantings using red ants, Oecophylla smaragdina.A
consultant report to the Livestock Development Corporation, Port
Moresby, Papua New Guinea, March, 2001.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 1995. The effect of the green ant,
Oecophylla smaragdina (Hymenoptera: Formicidae), on insect pests of
cashew trees in Australia. Bulletin of Entomological Research, 85, 279 –
284.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 1998a. How many queens are
there in established green ant, Oecophylla smaragdina, colonies? The
Australian Journal of Entomology, 37, 249 – 253.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 1998b. Locating queen ant nests in
Oecophylla maragdina (F.) colonies. Insectes Sociaux, 45, 477 – 480.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 1999. The effect of colony isolation
of the predacious ant, Oecophylla smaragdina (F.) (Hymenoptera:
Formicidae), on protection of cashew plantations from insect pests.
International Journal of Pest Management, 45, 189 – 194.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 2000a. Why use weaver ants,
Oecophylla spp, to manage the main insect pests in cashew orchards? In
‘‘Cashing in on Cashew’’ International Workshop on cashew production
and processing, Marawila, Sri Lanka, 13 – 15 November, 2000.
PENG, R. K., CHRISTIAN, K. and GIBB, K., 2000b. Continuation of
transplanted green ant colony monitoring and implementation of ant
technology in commercial cashew plantations. A research report for the
Rural Industries Research and Development Corporation, Canberra,
Australia, December, 2000.
POFFLEY, M., 1996. Mango management—flowering to market. Department of
Primary Industries and Fisheries, Northern Territory, Australia, Agnote, No.
301, September, 1996.
SIEGEL, S., 1956. Nonparametric statistics for the behavioural sciences (New
York: McGraw-Hill Book Company, Inc.).
VAN DER PLANK, F. L., 1960. The bionomics and ecology of the red tree ant,
Oecophylla sp., and its relationship to the coconut bug Pseudotheraptus
wayi Brown (Coreidae). Journal of Animal Ecology, 29, 15 – 33.
VAN MELE, P. and CUC, N. T. T., 2000. Evolution and status of Oecophylla
smaragdina as a pest control agent in citrus in the Mekong Delta, Vietnam.
International Journal of Pest Management, 46, 295 – 301.
VAN MELE, P. and CUC, N. T. T., 2003. Ants as friends: Improving your tree
crops with weaver ants. UK: CABI Bioscience, pp.66.
VAN MELE, P., CUC, N. T. T. and VAN HUIS, A., 2002. Direct and indirect
influence of the weaver ant Oecophylla smaragdina on citrus farmer’s pest
perceptions and management practices in Mekong Delta, Vietnam.
International Journal of Pest Management, 48, 225 – 232.
WAY, M. J., 1954. Studies of the life history and ecology of the ant Oecophylla
longinoda Latreille. Bulletin of Entomological Research, 45, 93 – 112.
WAY, M. J., 1963. Mutualism between ants and honeydew-producing
Homoptera. Annual Review of Entomology, 8, 307 – 344.
WAY, M. J. and KHOO, K. C., 1992. Role of ants in pest management. Annual
Review of Entomology, 37, 479 – 503.
WILKINSON, L., 1990. SYSTAT: The system for statistics. (Evanston, IL:
SYSTAT Inc).
YOUNG, G. and CHIN, D., 1998. Redbanded thrips on fruit trees. Department of
Primary Industries and Fisheries, Northern Territory, Australia, Agnote No.
752, May, 1998.
YOUNG, G. and POFFLEY, M., 1997. Red-banded thrips (Selenothrips
rubrocinctus). Department of Primary Industries and Fisheries, Northern
Territory, Australia, Agnote No. 719, August, 1997.
114 R. K. Peng and K. Christian
