Dragonfly larvae are generalised predators, able to attack, overcome, and consume a tremendous diversity of potential prey types. Typically prey varies from tiny crustaceans to fish and tadpoles. However, not all dragonfly larvae can, or do, attack all types of prey. There are two underlying causes for this. Dragonfly larvae tend to be microhabitat specialists, which has a major effect on determining the prey they are likely to meet. Also the food-catching structure (the labium or "mask") in many groups has become specialised, increasing the efficiency of attacks on certain kinds of prey, but making other kinds of prey immune from attack. However, even larvae with quite specialized structures may still capture and consume unexpected prey types when an opportunity arises.
A. labium of a coenagrionid damselfly - a generalized form
B. labium of a Lestes species damselfly - specialized for a long reach and a fast attack on small prey
C. labium of a calopterygid damselfly - catches large, slow-moving insects (Plecoptera, Diptera identified)
D. labium of a petalurid dragonfly - a generalized, heavy duty form, often used to attack large prey
E. labium of a libellulid dragonfly - the setae stand up during the strike forming a basket, catches small and mobile prey.
When describing the features of the labium scientists usually use a terminology developed by Corbet (Corbet, Philip S. 1953 A terminology for the labium of larval Odonata. Entomologist 86: 191-6.).
The "economic" effect of dragonfly larval predation is poorly understood. Most studies have looked only at the impact of large dragonfly larvae on larger stages of prey types sharing the habitat (animals which presumably have survived in the face of any predatory pressure). The effects of the smaller larvae and the younger phases of prospective prey also need to be taken into account. Predation by dragonfly larvae can prevent species from using biologically suitable habitats so the presence is unmeasurable. Documented examples of this include fish farms and frog ponds where dragonfly larvae have eliminated fish hatchlings and small tadpoles, leaving nothing. As a thought experiment contrast the density of mosquito larvae in an old tyre with that in an adjacent tiny pond with a few damselfly larvae ?