SCIENTIFIC CORRESPONDENCE A fly's ultimate con SIR - Many aberrant insects are known females whose wings are either b ~ from colonies of social insects 1, but few as absent or reduced, or whose ,\ . - E. bizarre as that reported here. In the rain­ membranes are shed when the fly forest of Malaysia, in 1994, two of us enters its host's nest. They are C2-....( (AW. and U.M.) collected an entire normally transported by the bivouac of a driver ant, an undescribed males to the colonies of their hosts species of Aenictus Shuckard, situated during mating2• All these aberrant in the fork of a tree. It was formed by phorids possess legs. Our speci­ ants that had interlocked their tarsal mens are all mature females, claws to form a temporary nest-like gravid with eggs (Fig. 2a). These structure. From this bivouac, one of us eggs are also examples of mimicry, (K.R.) recovered 104 specimens of an closely resembling those of their unidentified insect (Figs 1, 2), 80 Diptera ant host. The presence of sperm, in larvae of the family Phoridae, 57,100 a spermathecal pocket discharging worker ants and a single queen. The latter into the atrium, indicates that the was in a non-physogastric state, indicating species has males. that the colony was in its migratory phase. Vestigial legs are previously The unidentified insects were flightless, unrecorded in Phoridae, and arc legless ( apart from reduced coxae exceedingly rare in adult Diptera. and trochanters), and had vermiform In the bat parasites Ascodipteron abdomens lacking tergites 2-7, but bearing Adensamer (Streblidae ), the newly strong setae. Despite the larviform appear­ emerged females possess both ance, the structure of the head and the wings and legs, but, on invading abdominal terminalia indicated that they their hosts, they shed the wings and FIG. 2 Legless fly and its leg rudiments compared with 3 are adults which closely resemble the the legs beyond their trochanters . those of a normal phorid. a, Whole fly showing mature larvae of their ant host. The form of the The closest parallel to our bizarre eggs within. b , c. Middle and hind coxae of legless proboscis, antennae and thorax (Fig. la) phorid is the marine midge phorid fly (b) and Metopina o/igoneura (c) . C2, middle immediately excluded them from Hymen­ Pontomyia Edwards (Chironomi­ coxae; C3, hind coxae; E, endoskeleton (including meso• optera and identified them as Diptera. dae ), whose females possess furca to left and metafurca to right); F2 , base of mid After failing to match these flies to any vermiform abdomens and lack femur; F3, base of hind femur. Scale bars, 0.1 mm. known myrmecophilous genera, a sample wings, halteres and front legs at the was sent to David Kistner (California time of emergence. Their middle and hind where7• Evidently, the flies and their State University, Chico), who, in collab­ legs are vestigial, being reduced to coxae, larvae are fully integrated into the ant oration with A. and M. Newton (Field trochanters and, in some cases, abbreviated colony and totally deceive their hosts. Both Museum of Natural History, Chicago), femora as well. In one species there are must be producing pheromones that mimic concluded that they might be highly also vestigial tibiae45• those of their host. In standard keys to aberrant Phoridae. This was confirmed by In our phorid, the middle and hind insect orders8, the lack of legs would cause one of us (R.H.L.D.) from the structure of coxae are the most developed, but the these adult flies to be identified the antennae and clypeo-cibarium. endoskeleton associated with these coxae as anomalous larvae of Diptera or other Many myrmecophilous and termi­ is not fully developed. Thus, the metafurca endopterygote insects. It is probable, tophilous Phoridae possess flightless is not fully sclerotized ventrally and lacks therefore, that others have collected any sclerotized connections with these highly aberrant flies, but have been the mesofurca (Fig. 2h). Such an unsuccessful in identifying them*. incomplete endoskeleton would be A. Weissflog unlikely to allow proper function­ U. Maschwltz ing of the legs if they were still Zoologisches lnstitut, present. Similarly, the exoskeleton Johann Wolfgang Goethe-Universitat, of the thorax shows incomplete 60054 Frankfurt am Main, Germany sclerotization and seems to be R. H. L. Disney insufficiently fused to allow University Department of Zoology, normal flight or walking. Thus, University of Cambridge, the mesonotal and metanotal Cambridge CB2 3EJ, UK phragmata are not fused, so that K. Rosciszewski the latter appears detached. The Staatliches Museum fur Naturkunde, only representatives of wings Karlsruhe, D-76042 Karlsruhe, Germany are the tegulae, basicostas and 1. Kistner, D. H. in Social Insects Vol. 3 (ed. Hermann, H. roots of the costas (Fig. la). The R.) 1-244 (Academic, New York, 1982). halteres are vestigial, with 2. Disney, R. H. L. Scuttle Ries: The Phoridae (Chapman & Hall, London, 1994 ). uninflated knobs. These reduced 3. Muir, F. Bull. Mus. comp. Zoo/. 54, 34 9- 366 (1912). structures represent a striking case 4. Edwards, F. W. Proc. zoo/. Soc. Lond. 51/ 52, 779- 806 (1926). of heterochrony, a phenomenon 5. Tokunaga, M. Mem. Coll. Agric. Kyoto Univ. 19, 1- 56 associated with other aberrant (1932). Phoridae and most strongly 6. Disney, R. H. L. Bonn. zoo/. Beitr. (in the press). 7. Disney, R.H. L. Sociobio/ogy(in the press). manifested by some of the 8. Tilling, S.M. Fld Stud. 6, 695-766 (1987). termitophilous Termitoxeniinae2.o. *If others have collected similar legless flies but failed to FIG. 1 Left face of legless fly. a, Head and t horax; This phorid will be assigned to a identify them to their order or family, R.H.L.D. would be b, whole fly. new genus and fully described else- pleased to examine such specimens. NATURE · VOL 378 · 9 NOVEMBER 1995 137