Female ovipositor

Definition: 

The ovipositor of acalyptrate flies consists of segments 6 to 8 and paired cerci, the latter being appendages of segment 11.

Character evolution: 

ovipositor, dorsal and ventral viewIn the stem-species pattern of the Chloropidae family-group and of the Milichiidae, the ovipositor is long and retractable. I suggest that all the sclerites of the ovipositor were originally complete, and that divisions or losses of sclerites have occurred several times independently. I have come to this conclusion because in all four families of the Chloropidaeovipositor, dorsal and ventral view family-group there are species with complete sclerites as well as species with longitudinally divided or absent sclerites. It is more probable to assume that sclerites were divided or lost several times, rather than that they are secondarily undivided or secondarily present. The same holds true for the evolution of a ring of barbed spines between segments 7 and 8: this ring can be found in some species in each family of the Chloropidae family-group as well as in several other families of acalyptrate Diptera, and must either have evolved repeatedly or have been lost several times independently. I suggest that the ring of barbed spines is part of the stem-species pattern of the Milichiidae and of the stem-species pattern of the Chloropidae family-group and that it was lost several times independently.
Within the Milichiidae, the plesiomorphic state of the ovipositor sclerites (all sclerites undivided) is present only in Borneomyia and in Leptometopa niveipennis. More frequently, S8 is modified or absent, and in addition S7, T7, or T8 are sometimes divided. The plesiomorphic state of the ring of ovipositor, dorsal, ventral and lateral viewsmall barbed spines between segments 7 and 8 is only present in Enigmilichia dimorphica, Eusiphona sp. 1, some Leptometopa species, and Madiza glabra. In the Phyllomyzinae, in the Leptometopa niveipennis-group, and in several Milichiinae the ring is lost. In contrast to this, the barbed spines are strongly developed in Desmometopa and Pholeomyia.
In Desmometopa and Pholeomyia, the ovipositor is short, with T6-7 and S6-7 undivided, S8 specialised, and T8 considerably shortened. The shape of S6 is triangular in Desmometopa and Pholeomyia and has a median projection on the anterior margin in Desmometopa(Litometopa). ovipositor, lateral view, SEMInside segment 7, females have a specialised structure which is called the secondary ovipositor. At oviposition the secondary ovipositor is everted telescope-like and then lies between S7 and the subanal plate. Basally the structure consists of a broad membraneous ring with many anteriorly-directed barbed spines. This ring probably originated from the ring of spines between segments 7 and 8 in the stem-species pattern of the Milichiidae. Distally there are two weakly sclerotised strips, each with a row of setulae. These strips originate ventrolaterally and meet dorsally at the functional tip of the ovipositor. Between these strips, which represent S8, lies the genital opening. When the secondary ovipositor is retracted, only the basal part is turned over, with the barbed spines now directed posteriorly. In addition to the secondary ovipositor, Desmometopa and Pholeomyia are characterised by a bare supra-anal plate and secondary ovipositor, indrawnshort, round cerci. This is in contrast to the stem-species pattern of the Milichiidae and of the Chloropidae family-group, where there is one pair of setulae on the supra-anal plate and the cerci are elongate.
The close similarity of the complex-structured secondary ovipositor in Desmometopa and Pholeomyia suggests that it is a homologous structure and evolved only once. As a consequence, the secondary ovipositor must have evolved in the stem-species of the Madizinae+Milichiinae and was reduced once or twice in the Madizinae and several times in the Milichiinae.
A secondary ovipositor occurs only in the Milichiidae, but similar, though not homologous, structures can be found in e.g. the Agromyzidae (Spencer 1987) and the Drosophilidae (Grimaldi 1987). In the Agromyzidae, segment 7 is modified into a non-retractable oviscapt into which the following segments can be retracted. The membrane between segments 7 and 8 bears anteriorly pointed denticles of different sizes and shapes, which are used to tear open plant tissue. S8 is developed into paired "egg guides".
The function of the secondary ovipositor in Desmometopa and Pholeomyia is not known. The resemblance between the secondary ovipositor and the ovipositor of plant-mining Agromyzidae may indicate that the females of these milichiid genera lay their eggs into relatively fresh plant material, which decomposes afterwards. Several Desmometopa species have been bred from decomposing plant material and larvae of Pholeomyia have been found in waste chambers of leaf-cutting ants. The egg-laying behaviour of Milichiidae is completely unknown. The strong, barbed spines may indicate that the females hook up on the substrate. This would make it easier for the female to push the ovipositor into more solid substrates. The short, and with its complete sclerites strong ovipositor of the species in question indicates a more solid substrate for egg-laying as well.
In the Milichia speciosa-group there are species with a short ovipositor (Milichia speciosa, Milichia sp. 1) which is very similar to the ovipositor of Pholeomyia: T6-8 and S6-7 are undivided, the anterior margin of S6 is convex, S8 is divided into narrow strips, the pair of setulae on the supra-anal plate is absent, and the cerci are short and round. However, the ovipositor of these Milichia species does not have the secondary ovipositor. There are also Milichia species with a long ovipositor, for example Milichia sp. 2. I believe that a bare supra-anal plate and short cerci are part of the character complex of the secondary ovipositor, and so the shape of the ovipositor in the Milichia speciosa-group is probably a transitional stage between an ovipositor with secondary ovipositor and one in which it has been lost.
In Leptometopa the degree of sclerite subdivision in the ovipositor differs strongly within the genus. L. niveipennis has undivided sclerites, whereas L. latipes and L. rufifrons, for example, have longitudinally divided T7, T8, and S8, as well as more or less completely divided S6 and S7. The Leptometopa niveipennis-group and the L. latipes-group, to which L. rufifrons belongs, differ in other characters as well (Papp 1984), and the monophyly of the genus needs to be checked. (ex Brake 2000)

Explanation: 

Inserted between segment 8 and the cerci is another section, the sclerites of which have not yet been satisfactorily homologised. These sclerites are called S9 and T9 by Matsuda (1976), hypoproct and epiproct (consisting of elements from segments 9 and 10) by J. F. McAlpine (1981), or sub- and supra-anal plate by Andersson (1977). I am using the terms sub- and supra-anal plate, because they describe the relative positions of the sclerites and do not suggest any homologisation with the sclerites of a particular segment. The female genital opening lies behind S8, and the anal opening between subanal plate and cerci. (ex Brake 2000)

Scratchpads developed and conceived by (alphabetical): Ed Baker, Katherine Bouton Alice Heaton Dimitris Koureas, Laurence Livermore, Dave Roberts, Simon Rycroft, Ben Scott, Vince Smith