The epiphallus is a narrow bridge connecting two lateral plates which bear small hooked ancorae on their anterior margins, and broad, rounded, widely separated, lobe-like lophi on their posterior margins. Lateral oval sclerites are present, roughly triangular in shape in most species, though oblong and strap-like in A. lemarineli. The epiphallic membrane runs backwards from the epiphallus to anchor on the zygoma and the edges of the ectophallic rami; there are no post-epiphallic sclerites, but in some individuals there is a weak oblong thickening of the membrane ventrolaterally, anterior to the the cingular rami.

The cingulum has a well-sclerotised zygoma, which bears short, curved cingular apodemes, giving a horseshoe shaped appearance in dorsal view (these are distinctly different from the longer, more orthogonally arranged apodemes of Serpusia – see below). The posterior margin of the zygoma forms a projecting ridge, seen as a short beaked projection in lateral view. The ectophallic rami are well developed, roughly oblong in lateral view; they extend to the ventral midline of the phallus, and meet but do not fuse at their ventral extremities. From their posterior margins an ectophallic sheath arises, forming two closely juxtaposed structures: a) a thick semitransparent membrane that covers the aedeagal valves ventro-laterally as they emerge from between the rami, and b) an opaque reinforced ectophallic sheath which envelops the tips of the aedeagal valves. The arch of the cingulum is well developed and forms a saddle-like sclerite on the posterior end of the spermatophore sac below the zygoma. It is rigidly connected to the inner surface of the zygoma by two oblique columns, which have to be cut through to extract the endophallus from the cingulum. Ventrolaterally, there are also connections between the arch and the inner surfaces of the rami. The cingular valves are short and deep, and at their extremities are fused with the inner surface of the ectophallic aedeagal sheath into which they protrude. The anterior margins of the rami bear a fringe of filamentous chitinous material; this fringing material runs from the ventral margin of the ramus upwards along its anterior margin to extend along the ventral edge of the cingular apodemes. A similar fringe is present on the dorsal edge of the proximal endophallic sclerites, and seems to be characteristic of the genus; we do not recall seeing it in other genera.

The anterior endophallic apodemes are short and concave, widely separated and arranged at nearly 90° to the long axis of the phallus. The gonopore processes are long and thin, and run backwards to end nearly halfway to the flexure. The ejaculatory sac is large and capacious, roughly spherical. The dorsal surface of the basal endophallic sclerites is “ornamented” with a chitinous fringe, as mentioned in the preceding paragraph. The flexure is fine and flexible, S-shaped in lateral view, and connects to the endophallic valves, which are simple slender curved sclerites running beneath the arch and inserting into the ectophallic aedeagal sheath. Like the cingular valves, their tips do not lie free, but instead are fused to the inner wall of the aedeagal sheath. In this genus it seems that the functional aedeagal valves are composed primarily of ectophallic sheath material, rather than of the actual endo- and ectophallic sclerites, which are incorporated into the lateral walls of the sheath.

The complexity of the phallic anatomy within the Serpusiae, and especially of the aedeagus, was previously examined by Jago (1970), in his study of the genus Auloserpusia. To elucidate the aedeagal structure he was forced to use the technique of serial microscope sections, and presented serial transverse sections of the aedeagal region of Auloserpusia poecila Jago, 1970 and remarked on its complexity. He wrote “Penis apex therefore consists of a triple sheath: a core of median penis valve and cingular valve, surrounded in turn by sheath derived from ventral side of penis valves and an outer double membranous part attached to posterior side of cingular arch. Penis valves often attached by cuticular material to the outer sheath”.

His diagrams support our conclusions in the above description, including the fusion of the aedeagal sclerites with the surrounding ectophallic sheath, and the connections between the arch and the cingular rami. He considered some elements of the aedeagal sheath to be derived from lateral lobes of the arch. We were unable to confirm this with simple dissection techniques.

The genera Serpusia and Aresceutica are essentially indistinguishable in their outer morphology. Jago (1970) wrote of them “both the above genera must be defined together” and gave only a single morphological description covering both genera. We find that only the prosternal tubercle separates the two genera reliably. In both species of Serpusia this is short, acutely conical and vertical, whereas in all species of Aresceutica it is longer, subconical, antero-posteriorly compressed, and tilted somewhat backwards. We therefore examined the phallic anatomy to see whether these are more discriminatory.

The phallic complexes of S. opacula Karsch, 1891 (type species of the genus) and of S. succursor Karsch,1896 are shown in Figs 7, 8. These are the two species said by Jago (1970) to be “true” Serpusia (see Introduction), and they have very similar phalli, close to those of Aresceutica, but differing from the latter in a number of details:

Figure 7. 

Serpusia opacula, phallus. A. Epiphallus, axial view; B. Epiphallus, dorsal view; C. Phallic complex after removal of epiphallus and epiphallic membrane, lateral view; D. As in C, but dorsal view. In this preparation the cingular apodemes are broken off; intact, they resemble those of S. succursor (Fig. 8), being long and straight and almost parallel; E. Endophallus and arch, lateral view; F. As in E, but dorsal view. In E and F, the arrows indicate cut margins of the arch where it has been freed from the inner surface of the zygoma.

1. The complex as a whole is more elongate and less robust than that of Aresceutica.

2. The endophallic apodemes are more deeply cupped, and from their ventral margins two channels with a U-shaped cross section run ventrally and posteriorly to form the gonopore processes. The gonopore processes of Aresceutica lack this channeled profile, and are longer and more slender, extending further rearwards. As a consequence, the ejaculatory sac, which is bounded by the gonopore processes, is situated more anteriorly in Serpusia than in Aresceutica, where it is distinctly ventral in position under the basal endophallus.

3. The flexure of the endophallus is very slender in Aresceutica, but in Serpusia it is thread-like and quite difficult to see at all if the preparation is not stained to visualise the sclerifications.

4. The cingular apodemes of Serpusia are long and more or less straight, running roughly parallel to each other distally. They are not curved to form a horseshoe-like profile in dorsal view as in Aresceutica (compare Fig. 8C with Fig. 4C or Fig. 5E).

5. The anterior margin of the cingular arch under the zygoma is broadly emarginate in the midline in Serpusia, whereas that of Aresceutica is straight.

6. The cingular valves are slender and pointed, not short and laterally flattened as in Aresceutica. They run along the upper margin of the endophallic valves, but end at the level where the latter enter the aedeagal sheath.

7. The endophallic valves are long and slender, wide and dorsoventrally flattened basally, but laterally compressed and narrow apically. They extend almost to the tip of the ectophallic aedeagal sheath, and are the major component of the aedeagus.

8. A complex ectophallic aedeagal sheath envelops the aedeagal valves as in Aresceutica, but is much less robust and is closely appressed to the endophallic valves, to which it is attached but not so extensively fused as in Aresceutica. In the intact phallus the projecting aedeagus of Serpusia therefore looks much more slender than that of Aresceutica, due to the wider ectophallic sheath of the latter.

9. The epiphalli of the two genera are similar, but the ancorae of Serpusia are stronger and are slanted inwards towards the mid-line, whereas those of Aresceutica point directly forwards (compare Figs 6B and 3B with Figs 7B and 8B).

Figure 8. 

Serpusia succursor, phallus. A. Epiphallus, axial view; B. Epiphallus, dorsal view. C. Phallic complex after removal of epiphallus and epiphallic membrane, dorsal view. The two cross hatched areas are the zones of attachment of the the arch to the inner surface of the zygoma; D. As in C, but lateral view; E. Endophallus and arch in lateral view; F. As in E but dorsal view. The arrows indicate cut margins of the arch where it has been freed from the inner surface of the zygoma; G. As F, but ventral view. In E, F and G the ejaculatory sac has been removed to show the detail of the gonopore processes.

Kevan (1956:20) described Aresceutica vansomereni from the Aberdare Mts. of Kenya (holotype male, Kenya, Aberdare Mts., Katamayu; in BMNH London). This species differs from the other members of Aresceutica in several external features, as was noted by Kevan in his description; he wrote “it differs in its stouter appearance, less strongly compressed and differently shaped prosternal tubercle, in the shape of the tegmina and in the coloration of the male”.

Examination of the phallus (Fig. 9) shows that it differs from those of Aresceutica species, but is similar to that described by Donskoff (1985) for Duviardia oubitai. This resemblance had been previously noted by Jago (unpublished MS). The external morphology of vansomereni fits the generic diagnosis of Duviardia given by Donskoff (1985) very well, allowing for the fact that the latter was based on only a single species (D. oubitai). The supranal plate is not trilobed at its extremity, but this feature is only very weakly seen even in the type species oubitai, as shown in Donskoff’s (1985) figure, and is absent in D. fishpooli Grunshaw, 1990. This character, together with the midline indentations of the anterior and posterior pronotal margins described by Donskoff (1985), seem to be specific to D. oubitai, and not properly part of the diagnosis of the genus.

The phallus of Duviardia vansomereni is closely similar to those of D. fishpooli and D. oubitai, and it differs from that of Aresceutica in several ways:

1. The aedeagus consists of two pairs of valves (ectophallic and endophallic), both of which are well separated and not fused to an ectophallic sheath at their tips;

2. The arch sclerite is connected to the inner surface of the zyoma by a single medial sclerification;

3. The gonopore processes are very short, causing them to be “éloigné de la flexure”, as described by Donskoff (1985).

4. The flexure of the endophallus is wide and robust, not thread-like.

5. The epiphallus is exactly like that figured for the other two species, and has considerably more robust ancorae than seen in Aresceutica.

On the basis of these observations, we formally transfer vansomereni from Aresceutica to Duviardia.

Figure 9. 

Duviardia vansomereni (comb. n.), phallus. A. Entire phallic complex, lateral view. Fine stippling indicates remaining portions of the epiphallic membrane; B. As in A, but dorsal view. The epiphallus is twisted out of its normal position in this preparation; C. As in A, but entire epiphallic layer removed; D. As in C, but after removal of ectophallic cingulum and rami, showing endophallus and ectophallic arch. The arrow indicates the cut edge of the arch sclerite. E, F, G. Epiphallus in axial, dorsal and lateral views.