Introduction
In most animal taxa the study of the male genitalia is of highest significance to solve taxonomic problems. That is also true for flies and leaf-mining flies in particular. Most agromyzid species cannot be identified without checking the genital characters. Frequently, species are found with completely different male genitalia but with virtually no external difference.
In the past there was much dispute about the biological and systematic meaning of the intriguing variability of genitalia (see for example Hennig, 1973, Griffiths, 1972 a, McAlpine, 1981, Sinclair, 2000). Sometimes it may be explained as pre-mating isolation mechanisms between closely related species ("lock and key hypothesis"). However, although intuitively convincing, this theory is not easy to confirm. Female choice and male competition that lead to permanent optimization can better explain the apparently rapid evolution of genital structures (further reading: Eberhard, 1985).

The genitalia of Agromyzidae are usually strongly developed and large. Unfortunately, the aedeagus (phallus), the taxonomic most important structure, is usually concealed below the last abdominal sternites at the ventral side of the abdomen (How to dissect and handle the male genitalia is described in the chapter Dissection of male genitalia). The living male can expand the aedeagus by means of the very strongly developed stick-like aedeagal apodeme. This organ can be considered as the most characteristic genital structure in the family Agromyzidae ( Ag nigrella genitalia.pct, Pseu europaea genitalia.pct). In some species it can be nearly as long as the whole abdomen.

In contrast to most previous taxonomic works I give here a broad overview of the different parts of the male genitalia, even if some of them are not of direct taxonomic use. I hope to improve the accessibility of the description by a number of digital images of male genitalia in situ.

Morphological terms

In the literature a lot of different terms for structures of the male genitalia in Diptera are used. This is because the nomenclature is based on competing hypotheses about homology. Except for some special terms applied for Agromyzidae only, in the following text mainly the terminology reviewed by Sinclair, 2000 is adopted. More about this subject can be read in Hennig, 1973, Griffiths, 1972 a, McAlpine, 1981, or Cummings et al., 1995.

The epandrium is the ninth segment of the male fly (Sinclair, 2000, McAlpine, 1981) Ag nigrella genitalia.pct. From posterodorsal view the epandrium of Agromyzidae is roughly horseshoe-shaped. Both the ends and the insides or the 'horseshoe' can be articulated and spread out as clasping organs. These structures are called surstyli. Phb cambii epandrium.pct They originate partially as lateral outgrowth of the epandrium and were partially formed by the bacilliform sclerites (Sinclair, 2000).
In most cases, where articulated surstyli are missing in agromyzid flies, they may have been fused with the epandrium since the inner sclerites of the epandrium -probably derived from the surstyli- normally are still present. This is the case in Agromyzinae and many Phytomyzinae, whereas in some Liriomyza species surstyli and inner sclerites apparently are completely lost.
Cerci are always present, situated in central position of the epandrium. They are sometimes conspicuously elongated (Agromyza, Ag nigrella genitalia.pct.), elongated and broadened (Japanagromyza, J inaequalis epandrium.pct) or short and often hyaline (e.g. Liriomyza Lir marginalis genitalia.pct).

The hypandrium (9th abdominal sternite) is situated mainly below and partly within the epandrium ( Ag nigrella genitalia.pct) where it is connected with the dorsal part. The easily visible part below the epandrium is triangular or 'u'-shaped. It consists of the hypandrial frame of different thickness, which often has some inner sclerotizations, where the gonites are attached. Chr syngenesiae hypandrium.pct, Oph pinguis hypandrium.pct.
About the non-sclerotized parts of the hypandrium not much is known since the male genitalia are normally studied in macerated condition.
Hypandrial apodeme is the term for the appendix of the hypandrial frame at the tip. It is widespread and sometimes particular pronounced within Japanagromyza ( J tristella aedeagus.pct, lateral view) and the Ophiomyia-group (Tribe Ophiomyiini) ( Hex simplex genitalia.pct, Kl pisi terminalia.pct).
In the latter picture the hypandrial apodeme is very distinct. However, that is not always the case, often an apodeme and a somewhat broadened hypandrial frame it can hardly be distinguished from each other.

The ejaculatory apodeme is also present in every agromyzid group. It is an unpaired sclerotized process of the sperm pump (ejaculator). Therefore it is connected with the aedeagus by a fine hyaline tube. The ejaculatory apodeme serves mainly as muscle attachment structure. Sometimes characters for this organ can be used for taxonomic and systematic purposes (see Genus Amauromyza).

Ag cinerascens genitalia.pct
Ag nigrociliata ej apodeme.pct
J desmodivora genitalia.pct
Oph lantanae ej apodeme.pct
Tr flacourtiae ej apodeme.pct

Am flavifrons genitalia.pct
Lir braziliensis ej apodeme.pct
Phb setosa ej apodeme.pct
Pseu europaea ej apodeme.pct


The gonites are normally connected at their thin, string-like base with the aedeagal apodeme and apically with the hypandrium. Their connection to the aedeagal apodeme can be tight and invariable as in Agromyza:

Ag albipennis genitalia.pct
Ag nigrella genitalia.pct

or loose and almost invisible as in other genera

Pseu europaea epandrium.pct
Pseu spicata epandrium v.pct

At their tip the gonites are immediately broadened and can have some peculiar stick- or leaf-like shapes as in Phytomyzinae ( Chr syngenesiae hypandrium.pct) and the Ophiomyia-group (Tribe Ophiomyiini): Oph pinguis hypandrium.pct

In previous literature two different kinds of gonites were considered: the pregonites and postgonites (Frick, 1952).
The latter are those structures here referred to as gonites. The pregonites were defined as the broadened areas at both sides of the hypandrium, where the respective (post)gonites are attached. Since the exact border between (pre)gonite and hypandrium is not visible, the term gonites is rejected here. It is of limited practical use.

The normally rather large and prominent aedeagus of male agromyzids shows an extreme variety of forms throughout the whole family. It is therefore widely used for species description and diagnosis.
In retracted position the aedeagus is inflated below the epandrium and the hypandrium and concealed by the last abdominal sternites. See the following pictures: Nap cichorii genitalia.pct, Lir sativae genitalia.pct.
For copulation the aedeagus comes out by means of the aedeagal apodeme, which is connected with the base of the aedeagus and the gonites. The rod-like aedeagal apodeme can be easily recognized and shows not much variation within Agromyzidae Chr nigra aedeagus full.pct. (In this figure the gonites and epandrium are removed.) The fully expanded position of the aedeagus may be illustrated with a picture of Phytomyza rufipes: Phy rufipes genitalia.pct

The aedeagal hood (epiphallus not exactly the same) Pseu europaea genitalia.pct, Lir bryoniae aed hood.pct (see also the previous figures) is the sclerotized distal part of the ejaculatory duct next to the attachment of the aedeagal apodeme. It is apically attached with the basiphallus and articulates with the aedeagal apodeme. The function of this organ becomes clear when it's appearance is compared in more retracted and in fully erected position:

Pseu europaea genit exp2.pct (aedeagus not fully erected)
Pseu europaea aedeagus exp2.pct (aedeagus fully erected)

In the retracted position of the aedeagus the aedeagal hood corresponds with its name and looks indeed like a hood at the base of the basiphallus. But if the aedeagal apodeme is pushed outside, the aedeagal hood will be pulled backward and tensioned. At the same time the aedeagus is stretched towards the body tip.

Although the shape of the aedeagal hood can vary significantly among the species, it was hardly used for diagnostic purpose. Exceptions are some Liriomyza species of economic importance (Liriomyza trifolii, Liriomyza huidobrensis, Lir bryoniae aed hood.pct, Liriomyza bryoniae). However, the use is limited.

To describe the aedeagus it can be primarily subdivided into two structurally different parts, the basiphallus and the distiphallus Ag nigrociliata aedeagus.pct. There are some suggestions for further subdivisions of the aedeagus (e.g. mesophallus: Frick, 1952, Nowakowski, 1962, Zlobin, 1994 a) but these terms cannot be unambiguously applied to a wide range of the agromyzid species. On this CD-ROM they are avoided. Instead, some merely descriptive names were tentatively applied to sections of the distiphallus.

The basiphallus consists of two side sclerites (Griffiths, 1972: paraphalli), which are usually asymmetrical or nearly symmetrical Ag nigrociliata aedeagus.pct. In certain groups (Ophiomyia, Tropicomyia) they can be somewhat twisted. Frequently the two basal side arms are subapically connected by one or two bridges ( Tr vigneae aedeagus dor.pct). This bridge can lead to the notion that the basiphallus is 'u'-shaped, a character used to recognize the genera Melanagromyza and Tropicomyia. In many Ophiomyia species the bridge between the side sclerites can be extremely oblique. When parts of the basiphallus are more or less hyaline it becomes difficult to distinguish between the bridge and one of the two side sclerites (illustrations: basiphallus mel oph.pct).
Sometimes apical filaments or appendices project from the aedeagus:

Ag albipennis genitalia.pct
Nem maculosa aedeagus.pct
Chr nigra aedeagus1.pct
Lir huidobrensis3 aedeagus.pct
Nap carotae aedeagus exp.pct

They can be of very different shape, size and origin. Especially those appendages located near the border between basiphallus and distiphallus normally can neither be homologized nor even assigned to one of the two subdivisions of the aedeagus.

Between the side sclerites the hyaline, usually more or less invisible ejaculatory duct is located.
The distiphallus is the enlarged, usually well pigmented apical part of the ejaculatory duct that has often a characteristically bent shape. Normally, the distiphallus ends in two terminal tubules, which can either be closely connected to each other or diverge apically (compare Ag nigrociliata aedeagus.pct and Phy ilicis aedeagus.pct). Sometimes the terminal tubules cannot be recognized as separate structures, e.g. in most Napomyza ( Nap carotae aedeagus exp.pct). The side arms of the basiphallus and the distiphallus are often overlapping because the basal part of the distiphallus can be situated between the side arms.
Basally the terminal tubules are branching from a single structure, the basal vesica of the distiphallus:

Ag nigrociliata aedeagus.pct
Pseu europaea aedeagus exp.pct
Lir huidobrensis aedeagus.pct
Tr flacourtiae genitalia.pct
Mel sojae aedeagus.pct

However, the basal vesica is missing or invisible in many agromyzid taxa.

The structural diversity of the aedeagus may be sometimes confusing. But usually a closer view reveals that in most cases a fair part of the observed pattern can be assigned to some of the structures described above, even if the whole aedeagus seems to be completely different to those examples depicted above.
For example in many Melanagromyza the basiphallus is hyaline and therefore more or less invisible ( Mel sojae aedeagus.pct). The short, peculiar, distiphallus is most characteristic for that genus. In contrast the terminal tubules of the distiphallus of certain Chromatomyia and Phytomyza species are considerably reduced Chr nigra aedeagus1.pct. Instead, the basiphallus and some appendages of it are most conspicuous and highly elaborate.