ECOLOGY & POPULATION BIOLOGY Ovipositor Length in a Guild of Parasitoids (Hymenoptera: Braconidae) Attacking Anastrepha spp. Fruit Flies (Diptera: Tephritidae) in Southern Mexico

In southernMexico, four native and one introduced species of Opiinae (Braconidae) attack larvae ofAnastrepha spp. fruit ßies. There is a substantial overlap in thehosts of the parasitoids, and every species has been collected from fruit ßies attacked by at least one or two other species. The ovipositors of these braconids have a broad interspeciÞc range of lengths, some are less than the length of the abdomen, and others, several times the length of the abdomen. The following three hypotheses are proposed to account for this variety of lengths: (1) Because of differences in the host stage attacked, there are differences in host vulnerability; i.e., mature host larvae feed at greater depths within fruit pulp and can be best reached with a longer ovipositor. There is an implication that competition among the wasp species has selected for foraging on different host stages and that thisdiversifying selectionhas resulted indifferentovipositor lengths. (2)Although longerovipositors increase host range and thus have competitive advantages, they may be heavy, awkward, and expensive to move around. If so, species with longer ovipositors might have to invest more in locomotion (reßected in wing size) and less in reproductive capacity (numbers of mature eggs held in the ovarial calyx). Balancing selection would then maintain both short ovipositor-small wingedhigh fecundity species and long ovipositor-large wing-low fecundity species. (3) Although there are niche overlaps among the species, each has a “core environment” determined by factors such as temperature, humidity, seasonality, and host diversity. Ovipositor lengths have evolved to met the requirements of these specialized environments, and are not due to interspeciÞc competition mediated by ovipositor length; i.e., there has been no diversifying or balancing selection for differences in ovipositor length. Hypothesis number 1 fails because all the species attack similar host stages.Neitherwas there support forhypothesis number2.Therewerenocorrelationsbetweenwing size, or potential fecundity, and ovipositor length. The lack of correlation between species-pairs niche overlaps and differences in species-pairs relative ovipositor length is most consistent with hypothesis number 3. That is, because species with similar ovipositors are neither more or less likely to co-occur in the same samples from various fruits and locations than species with different ovipositor lengths it may be that species interactions are unimportant in the evolution of ovipositor lengths. If so, the lack of a pattern of competition made easily recognizable by differences in ovipositor length could inßuence biological control tactics. For example, if it is difÞcult to predict the abilities of newly introduced species to integrate in the existing guild of natural enemies, it may be more prudent over the short term to concentrate on the conservation of the natural enemies already present rather than pursue the “classical” introductions of new species.