• No results found

Oecophylla longinoda (Hymenoptera: Formicidae) as a biological control agent for cocoa capsids (Hemiptera: Miridae)

N/A
N/A
Protected

Academic year: 2021

Share "Oecophylla longinoda (Hymenoptera: Formicidae) as a biological control agent for cocoa capsids (Hemiptera: Miridae)"

Copied!
10
0
0

Loading.... (view fulltext now)

Full text

(1)

P.M. van Wijngaarden, M. van Kessel & A. van Huis

Laboratory of Entomology, Wageningen University, PO Box 8031, 6700 EH Wageningen, The Netherlands, E-mail: machiel.vanwijngaarden@wur.nl

The biology and ecology of the ant Oecophylla longinoda(Hymenoptera: Formicidae) was studied in cocoa plantations in Ghana in 2005. The main focus was the possibility to use the ant as a biological control agent against capsids (Hemiptera: Miridae) which are important cocoa pests. The effectiveness of both red and black strain O. longinoda to control this pest was studied by relating its presence, abundance and number of nests to canopy defoliation and percentage of pods damaged. In total 300 cocoa trees in three different cocoa fields at the Cocoa Research Institute of Ghana premises were examined. Both the pres-ence of red O. longinodaand a high abundance of the ants were correlat-ed with a thick canopy. This might be the result of either decreascorrelat-ed defoliation (lower capsid damage through protection by the ants), a preference of the ants for thick canopies, or a relational combination of both. The percentage pod damage (24) in trees with high ant abundance was significantly lower than those in trees without ants (42). The num-ber of pods damaged was about 50% lower for trees with black strain O. longinodacompared to trees without ants. This reduction in capsid dam-age by black O. longinodamight be explained by its high aggressiveness. The ant and especially the black strain is therefore considered a poten-tial biological control agent of capsids. However, it seems that ant abundance need to be high such that ants are not only present in the tree canopy, but also cover the trunk where the vulnerable cocoa pods are situated. The ant’s role in an integrated pest management strategy should be further investigated, in particular its combination with the natural pesticide neem.

Keywords: Oecophylla longinoda, weaver ant, capsid damage, biological control, integrated pest management, Convergences of Sciences, Ghana, cocoa

The ‘Convergences of Sciences’ (CoS) project initiated from the idea that scien-tific innovations arerarely used by farmers and that its contribution to poverty alleviation is sub-optimal. It is believed that convergence is needed in

technolo-Formicidae) as a biological control agent for

cocoa capsids (Hemiptera: Miridae)

(2)

gy development, not only between natural and social scientists, but also between societal stakeholders (including farmers) and scientists. Of key importance within the CoS principle is that societal stakeholders influence scientists’ research agendas and vice versa (democratization of science), and also that farm-ers gain more confidence in their own knowledge and capacities (farmer

empow-erment) (Röling et al.2004). During the CoS project farmers came up with the

idea to use the weaver ant Oecophylla longinodaLatr. (Hymenoptera: Formicidae)

as a biological control agent against capsids (Hemiptera: Miridae) in cocoa

(Ayenor et al.2004). Ayenor (2006) showed that the capacity of the ants to

sup-press capsids was equally effective as the use of the botanical pesticide neem. Oecophylla longinoda(Hymenoptera: Formicidae) is the African variant of the Oecophyllafamily and occurs in most tropical areas of Africa. The ants inhabit tree canopies in which they build nests out of living leaves. Their feeding

behav-iour is generalistic (Way 1954, Vanderplank 1960) and Oecophyllais relatively

dominant compared to other ant species (Bigger 1981). Colonies might cover large areas, occupying a number of tree canopies. The ants inhabit several crops

such as cocoa, coffee and citrus (Van Mele & Cuc 2003). Oecophylla smaragdina

(Fabr.), the Asian congener of O. longinoda, has already been used for centuries

to protect crop harvests in South-East Asia (Van Mele & Cuc 2000). Since

sev-eral studies have shown significant reduced pest damage by O. smaragdinain for

example citrus plantations in Vietnam (as reviewed in Van Mele & Cuc 2003)

and mango (Peng & Christian 2004, 2005a, b) and cashew orchards (e.g.Peng et

al. 1999, 2001) in Australia, it is considered a good biological agent. Therefore,

this study focused on the possibility to use O. longinoda in an integrated pest

management strategy for cocoa.

Pods, the fruits of the cocoa tree containing cocoa beans, are mainly located at the trunk of the tree, from the base up to the canopy. Capsids attack these pods, which serve them as a source of food. The most abundant capsid species, Salbergella singularis(Hagl.), Distantella theobroma (Dist.) and Helopeltis species

(as reviewed in Brun et al.1997, Entwistle 1972), create lesions by injecting their

stylets into the husk of the pod. These lesions are by itself relatively harmless, but they mediate infection of the pods by pathogenic fungi. Of these fungi, Phytophthora megakaryais the most common one, causing the pod to die and turn

black (black pod disease) (Opoku et al. 2002). However, capsids not only attack

pods, but also feed on the young shoots called cocoa tree ‘chupons’. Infection of shoots and branches by fungi is believed to cause dieback, resulting in defoliat-ed trees (Entwistle 1972).

The purpose of this research was to give an alternative for the mass spraying of capsids with synthetic pesticides provided gratuitously by the Ghanaian gov-ernment. The aim is to contribute to a more environmental friendly way of pest suppression with virtually no costs. In addition, the price farmers get for the cocoa could increase, both because of better quality when pesticides are not used,

(3)

as the bonus companies are willing to pay for organic cocoa. We conducted a number of studies with the intention to enhance the effectiveness of the weaver ant in pest control. These studies involved issues such as the possibility to

intro-duce O. longinodanests, effectiveness of pest suppression, nest composition,

pop-ulation dynamics and food consumption. The main question was whether or not O. longinodapopulations would be able to reduce capsid damage. Therefore, an experiment was carried out to link ant presence and abundance with capsid dam-age.

MATERIALS AND METHODS

The experiments were carried out in three cocoa plantations at the Cocoa

Research Institute of Ghanapremises at Tafo (Ghana). All the plantations had not been sprayed with pesticides for the past season and no other experiments had

taken place which could have had an effect on both O. longinodaand capsid

pop-ulations. In six different periods between the 6thof June and the 24thof October

in 2005, 50 trees were selected for examination, resulting in a total of 300 trees. The trees were randomly chosen in plots of 10 by 10 trees.

Monitoring ant presence

Ant presence and abundance

For every selected tree the presence of the black or red strain of O. longinodawas

registered. Ant abundance was monitored by careful observation of the canopy

and trunk, using a ladder if necessary. Four scoring levels were used: no O.

longin-odapresent on the tree (0); few ants observed on only one or two branch(es) (1);

ants observed on several branches of the tree (2); and many ants observed on many branches of the tree and/or trunk (often many trails present) (3).

Nest counting

The number of nests per tree was recorded by carefully examining the tree canopy. Since in some cases it was not possible to determine the exact number of O. longinoda nests in the tree canopy (e.g. in cases of a very big or dense canopy) often an estimate of the number of nests was made, based on ant trails and ant abundance.

Monitoring capsid damage

Canopy

The Hammond index (Johnson & Burge 1971) is commonly used to estimate the percentage of canopy loss due to capsid damage. This index represents five lev-els for estimating canopy loss (0: no damage to the foliage, 1: 0-25% damage to

the foliage, 2: 26-50%, etc.). A variant involving 21 levels was designed to estimate

differences in canopy loss with a margin of less than 10%: no damage to the

(4)

Pods

All pods above the size of a human thumb were checked for presence of capsid lesions. Only pods located less than four meters from the ground were included. Pods located above this four meter level could not be reached for, and were there-fore not examined. The total number of pods was documented together with the number of pods with capsid lesions. With these data, the percentage of damaged pods per tree was calculated.

Chupons

It was recorded whether chupons showed capsid damage. These observations will be discussed.

Data analysis

The correlation between nest and ant abundance was tested with the Spearman test. To investigate ant-capsid correlations all variables were linked in either a univariate analysis (Mann-Whitney U test) or a multivariate analysis (Kruskal-Wallis). To determine interrelated differences, post hoc tests were run. All tests were carried out at the 0.05% confidence level.

Capsid presence and damage was also evaluated for different ant densities. A

divi-sion was made between red and black O. longinodato distinguish the effect of the

dif-ferent strains. To be able to compare the data of red and black O. longinodafor

simi-lar ant densities, an analysis was carried out only for trees with high ant abundance (index 3). At last, the number of ant nests was correlated with capsid damage. RESULTS

Canopy condition

There was no relation between ant presence in general and condition of tree canopies (Table 1). When canopy thickness is correlated with ant abundance, only trees with a high abundance of ants (index 3) had significantly higher leaf coverage compared to those with no ants (index 0). The abundance of the red ant strain is positively correlated with greater canopy thickness, whereas this does

not hold for the black strain of O. longinoda. A higher number of nests per tree

does not mean significantly less capsid damage. The nests were found to be in places with low canopy cover.

Pod damage

The presence of O. longinodais significantly correlated with lower pod damage

(Table 1, Fig. 1). Further, analyses show that pod damage is only significantly

lower when O. longinodais present in high numbers. When only trees with high

ant abundance (index 3) are compared to trees with no ants (index 0), the pod damage is significantly less (approximately 60%). However, a higher number of O. longinodanests is not significantly related to the level of pod damage.

(5)

Table 1.

Effect of the ant

Oecophylla

longinoda

on preventing capsid damage to the canopy and to pods in cocoa (average values)

Ant indices 1 Significant dif ference between 2 Canopy index Ant presence -9.1 (N=92) + 8.6 (N=208) MW : ns Ant abundance 0 9.1 (N=92) 1 8.8 (N=46) 2 8.9 (N=54) 3 8.3 (N=108) KW : 0 & 3 Ant strain 0 9.1 (N=92) R 8.3 (N=132) B 9.1 (N=72) KW : 0 & R

Ant strain (for abundance 3)

0 9.1 (N=92) R 7.9 (N=59) B 8.9 (N=45) KW : 0 & R Number of nests 0 9.1 (N=169) 1-4 8.0 (N=85) >4 7.6 (N=46) KW : 0 & 1-4; 0 & >4 Pods damaged (%) Ant presence -41.7 (N=68) + 28.7 (N=170) MW : – & + Ant abundance 0 42.1 (N=67) 1 39.2 (N=32) 2 30.9 (N=43) 3 24.0 (N=96) KW : 0 & 3 Ant strain 0 42.1 (N=67) R 32.6 (N=1 11) B 17.5 (N=56) KW : 0 & B; R & B

Ant strain (for abundance 3)

0 42.1 (N=67) R 25.9 (N=54) B 16.2 (N=38) KW : 0 & B; 0 & R Number of nests 0 35.8 (N=1 17) 1-4 26.9 (N=79) >4 33.4 (N=42) KW : ns

1Presence: – = absent, + = present; abundance index: 1, 2 and 3; strain: R = red, B = black, N = number of trees; 2P

0.05 (MW = Mann-Whitney U test,

KW = Kruskal-W

(6)

When a division is made between different strains, it appears that only for trees with black ants the pod damage is significantly lower (approximately 50%)

compared to trees without ants. Red O. longinodaants lower pod damage, but not

significantly.

Chupon damage

When O. longinoda was present in a tree, it was often observed that young

chupons had many ant trails on them. If a tree had many chupons and O.

longin-odawas present, the ants often nested in the chupon foliage. When O. longinoda

was present the chupons hardly showed any capsid damage. DISCUSSION

Canopy condition

The relation between ant occurrence and canopy thickness might have two

rea-sons. It is possible that trees with high O. longinodaabundance and high numbers

of nests have a thick (healthy) canopy because of reduced capsid damage by ants.

Another reason might be that O. longinodahas a preference for healthy canopies

for both environmental and ecological reasons. However, it is difficult to sort out these causal relationships. Another possibility is an interaction between both systems: a thick canopy attracts ants and due to the higher number of these pred-ators the canopy is better protected against capsids. This is a self reinforcing process: the canopy gets thicker as a result of reduced capsid damage and thus

attracts more O. longinoda, etc. Abundance of red O. longinodacorrelates better

with increased canopy coverage than abundance of the black strain does. We have no explanation for this difference between the two strains.

Figure 1.The effect of the ant Oecophylla longinodaon pod damage by capsids per cocoa

tree (%). A = Trees without, and with red or black ants. B = Trees without, and with red or black ants (no ants and abundance level 3 only). C = Trees with different ant abundances.

(7)

As ant abundance, the number of their nests is also related to canopy cover. Although the same explanation could be given as above, it is also likely that ants choose trees with a thick foliage as the habitat conditions are more favourable for nesting. This could be based on two reasons: either statistically there is a higher chance of finding a suitable nesting site in a thicker canopy, or they choose a thicker canopy to provide nesting sites for seasonal migration, which is suggested in the literature and by farmers (Way 1954, Vanderplank 1960). The fact that nests do not remain at one location during the year was supported by observations of colony movement for over more than three meters to other trees. Also farmers had depicted the seasonal movement of ants over their plantation. In addition to observations in literature (Way 1954, Vanderplank 1960) our own

observations confirm that nests of O. longinodawere often found on the outside

of the canopy in direct sunshine. Besides that, nests were observed in trees with a serious degree of defoliation.

Pod damage

Lower pod damage (approximately 50%) was observed in trees with black O.

longinoda present compared to trees without ants. This indicates that this ant strain could be a good biological control agent of capsids. However, it is not sure whether the same results would be achieved when replicated in cocoa fields hav-ing different environmental conditions.

Abundance of weaver ants might be very important in cocoa pest manage-ment. Pods mainly grow on the trunk and it is likely that these pods will only be protected against capsids if the ants also occur at the trunk and not only high up in the canopy, which seemed to be their preferred territory. The ant was only observed on the trunk of a tree if the ant population on that specific tree was high (index 2 or 3). If abundance was very low ants often only occurred in the top and/or middle part of the canopy. Therefore, in order to protect cocoa pods, O. longinoda populations have to be very abundant, whichis also suggested by

Vanderplank (1960). Other studies should focus on whether O. longinodacolonies

could protect a whole plantation in a sufficient manner. Naturally most weaver

ant colonies do only cover part of it (Peng & Christian 2005b). Black O.

longino-dacolonies were observed to vary in concentration: parts of the plantations were

highly occupied, where as other parts were not occupied at all. This could indi-cate that only parts of the plantation with a high ant occupation are protected.

The reason for the efficiency of black O. longinoda in controlling capsids

might be due to its aggressiveness. Several observations in the fields have shown that in case of disturbance or danger, black ants attack immediately whereas ants of the red strain tend to keep a distance. In case of encountering capsids, black ants always succeeded to catch the prey where as red ants often led it slip away. However, these are just observations and a study should be done in order to con-firm the reason for differences in efficiency. Besides, it might be important to know what exactly determines differences in strains, since it was sometimes

(8)

observed that the two strains were part of one and the same colony; also inter-mediate forms were found. It has been suggested that this is due to differences

in diet: red O. longinoda would often mainly feed on honeydew produced by

Homoptera; where as the black ants are more active predators (Vanderplank

1960). This would also explain why black O. longinodareduces capsid damage to

a greater extent than red strain populations.

Though not significant, the results indicate a decreasing percentage of dam-aged pods with an increasing number of ant nests.

Chupon preference

Observations in the field indicate that O. longinodaants possibly prefer chupons

above other branches in the canopy for nest building. This might be because chupons often have numerous big fresh leaves which possibly match the optimal conditions for nest building activities (Way 1954, Van Mele & Cuc 2003). Chupons are often very vulnerable for infestation of capsid lesions. Protection

of chupons would be an extra advantage of the use of O. longinodain cocoa pest

control.

Integrated pest management

The 50% reduction in capsid damage by black O. longinodais considerable, but it

may be questioned whether this is sufficient or not. Farmers perceive pesticides use to be effective. However, synthetic pesticides also wipe out ant colonies. It

remains to be seen whether O. longinodacan suppress capsids sufficiently, taking

into account that according to Entwistle (1972) only a few capsids can already have a devastating effect on cocoa trees. However, the method could be effective when used over a longer period. The method also does not pose any harm to the environment. Besides, it has low initial costs, does not require much labour, and

farmers are already convinced of O. longinoda’s usefulness.

Sporleder & Rapp (1998) showed that palm trees with O. longinodahad

signif-icantly higher numbers of coconuts (resulting in higher yield) than those with-out. In studies of Peng and his co-workers the number of nuts was positively

linked to O. longinoda population stability as well as fruit quality (Peng et al.

1998, 1999, Peng & Christian 2005b). It would be a valuable experiment to look at cocoa yield (such as number of pods, number of beans, bean weight, and

quali-ty) in relation to O. longinodapresence and population stability.

This study indicates that for successful biological pest control, O. longinoda

could be part of an integrated pest management strategy, including other compo-nents such as black pod removing, weeding and pruning. Therefore, ants should be introduced in plantations where they are absent and reinforced in plantations where ant abundance is low. Introduction of ants is an option (Van Mele & Cuc 2003), although other experiments within this research indicated some difficulties

like hostile native O. longinodacolonies (Peng et al.1999, Van Mele & Cuc 2003),

(9)

Peng & Christian 2005b) and failure of large scale introduction, which could also partly be the result of the first two reasons. Further research should focus on improvement of the introduction method and overcoming its basic problems.

Human intervention (so called ‘ant husbandry’) might be needed for O.

longino-da colonies to grow out to be an effective biological control agent. A detailed

description of introduction methods and ant husbandry is found in Van Mele & Cuc (2003). An addition to the IPM strategy including ant husbandry could be the use of the botanical pesticide neem. According to own experiments neem, applied

directly on or in the surroundings of O. longinoda nests or individuals, has no

direct deterrent effect or affect mortality rate. Neem was also included in the IPM strategy of the CoS project (Ayenor 2006, Dormon 2006).

The suggestion of using ants in pest control came from the farmers themselves. This increases the chance of adoption of the method by farmers. A few objections were raised, but all problems mentioned by farmers were easily resolved, even after discussions among themselves. What is needed is an easy, cost-effective and

non labour-intensive method of introducing and maintaining O. longinoda

popula-tions in cocoa plantapopula-tions. Besides that, the advice should be mediated to the farm-ers, for example through farmers fields schools (FFS). To facilitate this process, a leaflet with basic information accompanied the report of this study.

One of the greatest difficulties of introducing biological control using ants, are the free mass spraying activities. However, these mass sprayings are believed to stop in the near future (E.N.A. Dormon, personal communication). When farmers are forced to pay for capsid control, they will look for the most suitable

method, which should both be cheap and effective. The use of O. longinoda, if

necessary in combination with neem or other biological measurements, might provide the most convenient alternative.

REFERENCES

Ayenor, G.K. 2006. Capsid control for organic cocoa in Ghana: results of participatory learning and action research. [S.l.: s.n.]

Ayenor, G.K., Röling, N.G., Padi, B., Van Huis, A., Obeng-Ofori, D. & Atengdem, P.B. 2004. Converging farmers’ and scientists’ perspectives on researchable constraints on organic cocoa production in Ghana: results of a diagnostic study, NJAS – Wageningen Journal of Life Sciences52(3-4): 261-284

Bigger, M. 1981. Observations on the insect fauna of shaded and unshaded Amelonado cocoa. Bulletin of Entomological Research71: 107-119

Brun, L.A., Sounigo, O., Coulibaly, N. & Cilas, C. 1997. Methods of analysis for study-ing cocoa (Theobroma cacaoL.) susceptibility to mirids. Euphytica94(3): 349-359 Entwistle, P.F. 1972. Pests of Cocoa, Longman group limited,1stedition, pp. 130-132

Greenslade, P.J.M. 1972. Comparative ecology of four tropical ant species. Insectes Sociaux19: 195-212

Johnson, C.G. & Burge, G.A. 1971. Field trials of anti-capsid insecticides on farmers’ cocoa in Ghana, 1956-60. 2. Effects of different insecticides compared by counting capsids, and capsid-counting compared with counting the percentage of newly

(10)

dam-aged trees. Ghana Journal of Agricultural Science4(1): 33-38

Opoku, I.Y., Akrofi, A.Y. & Appiah, A.A. 2002. Shade trees are alternative hosts of the cocoa pathogen Phytophthora megakarya. Crop Protection21(8): 629-634

Peng, R.K. & Christian, K. 2004. 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. International Journal of Pest Management: 50(2): 107-114 Peng, R.K. & Christian, K. 2005a. The control efficacy of the weaver ant, Oecophylla

smaragdina(Hymenoptera: Formicidae), on the mango leafhopper, Idioscopus nitidu-lus (Hemiptera: Cicadellidae) in mango orchards in the Northern Territory. International Journal of Pest Management51(4): 297-304

Peng, R.K. & Christian, K. 2005b. Integrated pest management in mango orchards in the Northern

Territory Australia, using the weaver ant, Oecophylla smaragdina (Hymenoptera: Formicidae) as a key element. International Journal of Pest Management51(2): 149-155 Peng, R.K., Christian, K. & Gibb, K. 1998. Bioecology of Helopeltis pernicialis

(Heteroptera: Miridae) with particular reference to the control efficiency by Oecophylla smaragdina (Hymenoptera: Formicidae). Proceedings of the international cashew and coconut conference: trees for life – the key to development, Dar es Salaam, Tanzania, pp. 170-174

Peng, R.K., Christian, K. & 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 Management45(3): 189-194 Peng, R.K., Christian, K. & Gibb, K. 2001. Potential of using colonies of the green ant,

Oecophylla smaragdina (F.), to control cashew insect pests. Technical Bulletin Department of Primary Industry and Fisheries, Northern Territory of Australia288: 81-93 Röling, N.G., Hounkonnou, D., Offei, S.K., Toussou, R. & Van Huis, A. 2004. Linking science and farmers innovative capacity: diagnostic studies from Ghana and Benin. NJAS – Wageningen Journal of Life Sciences52: 211-235

Sporleder, M. & Rapp, G. 1998. The effect of Oecophylla longinoda (Latr.) (Hym., Formicidae) on coconut palm productivity with respect to Pseudotheraptus wayiBrown (Hem., Coreidae) damage in Zanzibar. Journal of Applied Entomology122(8): 475-481 Van Mele, P. & Cuc, N.T.T. 2000. Evolution and status of Oecophylla smaragdina

(Fabricius) as a pest control agent in citrus in the Mekong Delta, Vietnam. International Journal of Pest Management46: 295-301

Van Mele, P. & Cuc, N.T.T. 2003. Ants as friends: Improving your tree crops with weaver ants, CABI Publishing,1stedition

Vanderplank, F.L. 1960. The bionomics and ecology of the red tree ant, Oecophyllasp. and its relationship to the coconut bug Pseudotheraptus wayiBrown (Coreidae). Journal of Animal Ecology29: 15-33

Way, M.J. 1954. Studies of the life history and ecology of the ant Oecophylla longinoda Latreille. Bulletin of Entomological Research45: 93-112

Way, W.J. & Khoo, K.C. 1992. Colony dispersion and nesting habits of the ants, Dolichoderus Thoracicus and Oecophylla smaragdina (Hymenoptera: Formicidae), in relation to their success as biological control agents on cocoa. Bulletin of Entomological Research81: 341-350

Figure

Figure 1. The effect of the ant Oecophylla longinoda on pod damage by capsids per cocoa tree (%)

References

Related documents

Khi người phi công trên sâu khấu hát xong bài hát của mình thì bỗng nhiên từ phía cánh gà trái có một người đ{n ông mặc quần áo rằn ri chạy vào sân khấu, bắn

arrangements (either public or private) and attributes this to higher billing costs and extra expenses derived from non-exclusivity within the market area born by the firm

● Vote to Appoint Voting Member and Alternate to Represent School Committee at the MA Association of School Committee’s Annual Meeting, and Vote to direct them on how to vote

This paper adds to previous empirical literature by examining attributes of municipalities that influence the decision of local governments to engage in hybrid organizational

Before 2001, Italian law had required the provision of public water operations under one or more of the following organisational modes: a) direct municipal management, whereby

The British and Victorian experiences point to considerable challenges with transferring revenue risk. Twenty years later, South Australia’s government sought to re-introduce private

the Center in 2019; the euphoric, spirit-lifting global phenomenon Blue Man Group makes their Kennedy Center premiere; and Tony ® , Grammy ® , and Olivier ® –winning Best

This paper contributes to an understanding of competition in the German motor in- surance market, which is the most important line of business in the German property-