• Biocontrol, as a benign alternative to chemical control, should always be the preferred method of pest suppression.
  • Biocontrol should be attempted only if an appropriate method has been chosen and offers a sound prospect of being successful.
  • Augmentative release is the only biocontrol method in which release of dragonflies, as adults or larvae, carries any reasonable prospect of success, and this by its nature must be applied in a closed environment.
  • Living dragonflies should never be moved from one habitat to another except as part of an approved, and monitored, scientific project.

On dragonflies as biocontrol agents

The contribution entitled “Tips” that appeared on page 11 of WDA’s Agrion 4(1) implied that releasing adult dragonflies in nature could suppress biting flies, such as mosquitoes, as effectively as applying insecticide. For several reasons it is important that such a claim should not be allowed to pass without comment. First, there is almost certainly no secure evidence to support it; second, there are compelling theoretical arguments for supposing it to be unfounded; and third, because using dragonflies in this way is inappropriate, the resulting failure is likely to discredit the practice of biocontrol. I shall try to explain what I mean.

There are two approaches to biocontrol of pests: inoculation and inundation. Inoculation, sometimes called ‘classical biocontrol’, entails introducing natural enemies (parasites, parasitoids, pathogens or predators) of a pest into an environment where they are not already present. This approach, after the requisite precautions have been observed, can be feasible in situations where a pest has been introduced into a new country without its complement of natural enemies. If the inoculation is successful, the natural enemies multiply naturally until they reach a level such that they either eliminate the pest or keep the pest populations down to a level deemed acceptable to humans. Inoculation rarely succeeds, partly because damage thresholds recognised by humans are usually far lower than natural enemies can achieve (after all, an oligophagous predator needs to leave some prey to feed on), and partly because, if natural enemies attain high densities, either at the time of release or subsequently, they typically disperse, thus reducing their effectiveness for local suppression. The other approach to biocontrol is called inundation or augmentative release (AR). This entails prior estimation of the numbers of natural enemies needed (within a given area and a given time) to achieve suppression to the required level, and then releasing sufficient numbers into a closed system, i.e. an environment from which they cannot disperse. If the requisite conditions are satisfied, AR can be highly successful. Indeed it is used routinely in several countries for the suppression of pests in greenhouses – closed environments from which neither natural enemies nor pests can disperse.

The only example known to me of dragonflies having been used successfully for suppressing a pest insect was a case of AR. It entailed releasing known numbers of half-grown larvae of the libellulid Crocothemis servilia into domestic water-storage containers in Yangon (Rangoon) in Myanmar (Burma). The water-storage containers were being used by the aquatic stages of the Yellow Fever Mosquito Aedes aegypti, which was responsible for the transmission of Dengue Fever in that locality. More than 92% of the local population of A. aegypti was occupying the containers which, because of their function, were easily accessible to householders and control operators. The systematic release of dragonfly larvae during the monsoon season (the time when Dengue Fever was being transmitted by the mosquito) rapidly depressed the mosquito populations to a level lower than could have been achieved by any other known method, including treatment by chemical insecticide. The trials that demonstrated the effectiveness of this approach have been described by Sebastian et al. (1990). That account has been followed by several shorter papers devoted to explaining the distinction between the two different approaches to biocontrol, and enumerating the reasons why the Yangon trial was successful. I repeat these reasons here because they illustrate the requirements if a programme of AR using dragonflies is going to offer prospects of success.

The conditions that allowed the Yangon trial to be successful were:

  • the confinement of a very high proportion of the aquatic stages of the target species to containers accessible to control operators and householders;
  • the availability, when needed, of a predator able to exert control promptly;
  • a methodology for propagating the predator reliably and in sufficient numbers;
  • a means of distributing the predator among containers harbouring aquatic stages of the target species;
  • the desire by the local community to achieve suppression of the target species and to do so without using chemical insecticides in their domestic water supply; and
  • understanding and acceptance by the local community of the rationale of AR of predators in domestic water-storage containers and willingness to participate in practical aspects of the control programme.

These conditions, clearly validated in the Yangon trial, illustrate the essential distinction between inoculation and AR, and they show why careful research and planning need to underpin any serious attempt to use dragonflies for biocontrol. After the account of this work was published, it became evident that the distinction between the two approaches to biocontrol was not always being understood and this prompted me to publish several short articles (e.g. Kimminsia 1: 12 and 2: 13-14; Selysia 20: 3, 7) in which I tried to explain the distinction and thus emphasise the conditions that must obtain if dragonflies are to be used in biocontrol attempts. I have also analysed the Yangon trial in my recent book (1999: 119-121), listing the attributes that a dragonfly must possess if it is to serve successfully as a biocontrol agent, and suggesting environments in which biocontrol attempts might offer promise.

From time to time it has been suggested that collecting dragonfly larvae from one place and translocating them for rearing and release in another place, can effect pest suppression at their destination. From what I have said it will be clear that the chances of such an approach being effective are vanishingly small. Furthermore, such actions are, by a sharp irony, likely to be counterproductive, because translocating dragonfly larvae from one habitat (in which they have managed to survive the disproportionately high mortality associated with earlier instars) to another where conditions may be less suitable is likely to increase their mortality and thus reduce the overall effectiveness of the resulting adults in suppressing pest insects. Also, to move living organisms from one habitat to another should always be discouraged unless the translocation is part of a project approved by an authority with appropriate expertise and containing provision for rigorous pre- and post-release monitoring. Not only can the introduction of predators like dragonfly larvae disrupt the balance of the ecosystem that receives them, but such an operation can seriously undermine the work of biological recorders whose findings provide data essential for studies of distriution and biodiversity.

Literature cited

  • Corbet, P.S. (1999). Dragonflies. Behaviour and ecology of Odonata. Harley Books, Colchester.
  • Sebastian, A., Sein, M.M., Thu, M.M. & Corbet, P.S. (1990). Suppression of Aedes aegypti (L.) (Diptera: Culicidae) using augmentative release of dragonfly larvae (Odonata: Libellulidae) with community participation in Yangon, Myanmar. Bulletin of Entomological Research 89: 223-232.

Phillip Corbet, Crean Mill, Crean, St Buryan, Cornwall, UK
Published in Agrion, Volume 4(2), July 2000