To achieve control of a target weed using the classical biological control approach, self-sustaining populations of an agent must become established. We can only be confident an agent has become established after monitoring shows not only that the liberated insects have survived, but also that new generations have appeared.

Under some conditions there may be a lengthy period before we can be confident of establishment. If liberations are performed in a region with extremes of climate, the agent's population must also have survived the less favourable period before we can be confident of agent establishment.

If parasites or predators adapted to the agent are plentiful in the region, the agent may indeed be reproducing on the weed but actually declining in population, ultimately to the point of local extinction (an example of this is the Malaysian project against Mikania micrantha where establishment failed despite much effort and expenditure, presumably because of predators and/or parasitoids). Goeden and Louda (1976) reviewed biocontrol programs whose outcomes were influenced by natural enemies of agents following releases.

Little can be done about unpredictable misfortunes resulting from chance events. Probably every biocontrol worker has experienced losses of valuable agents, or even losses of entire release sites, by fires, floods, human mischief or a host of other causes. Mostly these are not preventable.

However, there is a great number of unfavourable factors or events which potentially can cause havoc to the efficient running of a biological control project and which are indeed preventable or avoidable.

The information on the pages to the left is not intended merely to give examples of what should or should not be done, but to also stimulate workers to consider what conditions and threats may be faced by agents following liberation. By anticipating problems and acting to avoid them or at least minimise their effects, the probability of successful establishment is surely enhanced.