Segment 9-11 of the male postabdomen
Because the discussion in Cumming, Sinclair & Wood is convincing and the descriptions and illustrations are readily comprehensible, I will follow the revised epandrial hypothesis on this website even though it has been criticised by some authors (Griffiths 1996, Zatwarnicki 1996).
According to Cumming, Sinclair & Wood (1995), the male genitalia in Schizophora consist of the dorsal saddle-shaped epandrium, which evolved from T9, and the ventral hypandrium, which evolved from a fusion of S9 and the gonocoxites. The tips of the U-shaped hypandrium articulate with the epandrium. Inside the epandrium this articulation is connected with the cerci and the surstyli via the subepandrial sclerite. The surstyli themselves articulate with the ventral side of the epandrium. The subepandrial sclerite evolved from a sclerotisation of the ventral membrane of the dorsal part of segment 9. The paired cerci belong to the 11th segment; between them lies the anus. The surstyli are secondary appendages of the epandrium. They hold the female ovipositor at copulation.
Inside the epandrium the base of the phallus, which is inserted into the female vagina at copulation, lies at rest. The phallus consists of the aedeagus fused with the parameres and is divided into basi- and distiphallus. At its base the phallus articulates with the rod-like phallapodeme, which is thought to have originated from a longitudinal invagination of the inner wall of the hypandrium. Laterally the basiphallus is encircled by the postgonites. Anterior to the postgonites are the pregonites, which are connected with the hypandrium anteriorly. Both postgonites and pregonites probably evolved from gonopods, but pregonites were newly formed within the Schizophora. The phallus is moved by the muscles of the phallapodeme; post- and pregonites prevent a lateral movement. 
From the phallus the ejaculatory duct leads to the sperm pump with the ejaculatory apodeme, which are situated shortly anterior to the epandrium inside the abdomen. The ejaculatory apodeme was originally a part of the aedeagus. The sperm pump consists of a membraneous sac, which is squeezed in order to press sperm into the female vagina. (ex Brake 2000)
The homologisation of structures in the male genitalia of Diptera has been a matter for considerable controversy: on the one hand, the homology with structures in other insects is unclear, and on the other hand the character transformation within Diptera is disputed. Basically, the hypotheses differ in the homologisation of the saddle-shaped dorsal sclerite (epandrium or periandrium), and its ventral appendages, the surstyli. The main theories are (1) the epandrial hypothesis of Crampton (1936), Chillcott (1958) and Hennig (1976); (2) the periandrial hypothesis of Griffiths (1972, 1981, 1991); and (3) the revised epandrial hypothesis of Cumming, Sinclair & Wood (1995). The third hypothesis is based on the first hypothesis. (ex Brake 2000)