Primitive Ghost Moths: Morphology and taxonomy of the Australian genus Fraus Walker (Lepidoptera: Hepialidae s. lat.)

By ES Nielsen and NP Kristensen

Hepialidae (ghost moths or swifts) are, in terms of diversity and distribution, the most successful group of homoneurous primitive moths. The morphology of Fraus is described in some detail with emphasis on the adult moth, and a new interpretation of hepialid male genitalia is presented. Beyond describing and illustrating a primitive hepialid, these observations are intended to serve as reference for the study of the classification of Hepialoidea and lower Lepidoptera.

In the taxonomic revision, based on more than 3000 specimens, the 25 Fraus species are described and diagnosed. The adult moths, as well as male and female genitalia, are richly illustrated, and distribution maps and flight period diagrams are provided for all species. The biology, behaviour, distribution and phylogeny are summarised and discussed.

• Language: English
• Publisher: CSIRO PUBLISHING
• Release date: Jan 1, 1989
• ISBN: 9780643105799

Book preview

Chapter 1

Morphology

We have not attempted to prepare an exhaustive morphological description of Fraus but have mainly paid attention to those body regions and organs which have so far been considered in comparative studies on lower Lepidoptera. Unless otherwise stated the observations on adult moths refer to species 1. minima. Integumental structure was examined in similar detail in 23. polyspila, and in cases where conditions in these two species (which represent, respectively, the lowest and the most advanced species-groups within the genus) are markedly different, representatives of additional species-groups have often been examined as well.

The Remarks sections are mostly concerned with comparative remarks on Fraus traits in a hepialoid context. Some of the morphological findings presented in this study are elsewhere (Nielsen and Kristensen in prep.) discussed in the broader context of the entire homoneurous-glossatan grade

Head

Head capsule

The head capsule (Figs 1-6) has a very low supraocular index, about 0.2 (1. minima) to below 0.1 (23. polyspila); the interocular index is about 1.4 (minima) to 1.9 (polyspila). The subgena does not extend below the compound eyes and no subgenal process is developed.

The antennal sockets are situated just below the level of the upper margins of the compound eyes; the dorsal and ventral cranial condyles are well developed. The antennal sockets are more or less extensively reinforced/melanized. In the examined members (20. fusca, 21. crocea, 22. simulans, 23. polyspila) of the simulans groups s. lat. there is a particularly strong reinforcement along the outer socket margin, from the ventral to the dorsal condyle.

There are few cephalic sutures (sulci). A short laterofacial suture may be formed as a linear extension of the antennal socket reinforcement, running towards the anterior tentorial pit. There is no frontoclypeal suture, but the scaled facial area (see below) must largely belong to the frons (and it is termed accordingly in the species descriptions) since the cibarial dilators insert below it, on the unscaled area. The postoccipital suture is consistently strongly developed along the dorsal and lower lateral (i.e. below the occipital condyles) foramen margins. The development of the remaining part of the postoccipital suture, as well as of temporal sutures, seems to be at least partly correlated with the development of concave bulges in the posterior cranial wall along the upper lateral margins of the foramen. Such bulges are well developed in the above mentioned members of the simulans groups s. lat., in which the adjacent foramen margins are strengthened by thickened postoccipital sutures; temporal sutures, thickenings arising from the dorsal foramen corners, similarly brace the upper margins of the bulges. In the examined members of the other species groups (1. minima, 10. serrata, 15. biloba and 17. nanus) the bulges, and one or both of the sutures, are less strongly developed. It is not easy, however, to describe this diversity in terms of clear-cut character states; indeed nanus is a somewhat intermediate form.

As always in the Exoporia, ocelli are lacking.

The cranial scale vestiture comprises a facial (frontal) patch situated on the somewhat bulging cranial area below the antennae and not reaching the eye, and a large dorsal patch from just behind/above the antennae to the hind margin of the vertex, and laterally continued downwards as a narrow band along the posterior and lower eye margin. An unsealed Y-shaped area (or a vestige of a Y) is present in the middle of the dorsal patch, probably indicating that there were once three discrete patches. A short row of scales between the antennal base and the eye is present in 22. simulans and 23. polyspila. Scale shape is diverse within the genus. Rather broad lamellar scales are present on the frons in males of most members of the bilineata group (Fig. 9) and the biloba group (Fig. 11); as a general rule, scales become narrower towards the top of the patch and females have narrower scales than males. The scales on the dorsal cranial area (‘vertex’) are relatively broad only in 1. minima (Fig. 8); the biloba group species have narrow lamellar scales together with piliform scales on the vertex, whereas the erect scales in other Fraus are piliform. In the pteromela and nanus groups lamellar frontal scales, when at all present, are quite narrow (Fig. 11), and in some pteromela group members and all simulans group s. lat. members (Fig. 12) all cranial scales are piliform.

Figs 1-6 F. minima, head. -1: anterior. - 2: posterior. - 3: ventral. - 4: dorsal. - 5: lateral. - 6: tentorium; ap: anterior tentorial pit; at: anterior tentorial arm; bm: basal maxillary piece; ct: corporotentorium; dc: dorsal cranial condyle; dt: dorsal tentorial arm; fc: frontoclypeus; lp: labial palp; md: mandible; mp: maxillary palp; oc: occipital condyle; pb: proboscis (galea); po: postlabial ‘wart’; pr: prelabium; tp: posteromedian tentorial process; vp: ventral process of anterior tentorial arm.

Figs 7, 8 F.minima, male head showing vestiture - 7: lateral - 8: dorsal.

Figs 9-12 Fraus species, heads showing vestiture. - 9: F. minima male, anterior -10: F. serrata male, anterior -11: F. biloba male, anterior - 12: F.polyspila female, anterior.

In 1. minima a pair of very small, blunt processes are present on the vertex; in 23. polyspila the corresponding formations are minute, so small as to be hardly discernible.

The tentorium has very prominent dorsal arms arising from dilated parts of the anterior arms. These dilated parts are sometimes (particularly in the simulans group s. lat.) markedly bent in the vertical plane and are bluntly produced downwards (Figs 6, 54); the ventral process is less developed in 17. nanus than in any other species examined. The anterior arms have a dorsally concave curvature behind the dorsal arms; they are united with the corporotentorium laterally, i.e. in the immediate vicinity of the posterior tentorial pits. The median process of the corporotentorium is well developed.

There is no discrete labrum in 1. minima or other primitive grade Fraus examined; the anteromedian part of the head capsule below the facial scale patch is strongly bent backwards and eventually it simply continues into the epipharyngeal membrane. A small medial lobe has been observed in 21. crocea and 22. simulans (Fig. 43) but not in 23. polyspila; its dorsal sclerotization is connected to the head capsule by lateral bridges originating adjacent to the mandibular articulation. The lower rim of the head capsule may be more strongly melanized than the adjacent areas; this is particularly true of the part just in front of the mandible vestige.

Remarks Out-group comparisons (with mnesarchaeoids and representatives of the smaller hepialoid ‘families’) lead us to believe that the presence of lamellar cranial scales is plesiomorphic in Fraus; scale narrowing has, then, apparently taken place independently in more lineages within the genus.

Tentorial shape in exoporian moths deserves more attention in a phylogenetic context: the ventral process of the anterior arm, usually so conspicuous in Fraus, has also been recorded from Afrotheora (Nielsen and Scoble 1986) and the Neotheoridae (Kristensen 1978a) but it is absent, or at most very faintly indicated, in the Mnesarchaeoidea (Kristensen 1968a and unpublished). The curvature of the anterior arm, similarly developed in Fraus and Afrotheora, also deviates from the mnesarchaeoid condition.

The absence of a differentiated labrum in primitive Fraus conforms with the condition reported from mnesarchaeoids, Anomoses, Neotheora, Afrotheora etc., and it might seem straightforward to assume that this is the exoporian groundplan state, and hence that the small labral lobes in some simulans group members are autapomorphic character reversals. Mortimer (1965) recorded the presence in Hepialus humuli of a similar small labral lobe, and we have confirmed this observation. The inferred character reversals therefore must have taken place on more than one occasion.

Compound eye

The compound eyes of Fraus are devoid of interommatidial setae.

Ommatidium structure is shown in Fig. 13. The corneal lenses lack internal processes. The crystalline cones are elongate, accounting for about one-quarter of the subcorneal eye height. The ‘fibrous retinal layer’ sensu Ehnbom (1948), i.e. the zone where the retinula cells are tapered and where most of their nuclei are located, is about as high as the cone layer; the nuclei are somewhat dispersed within the proximal part of this layer. The distal pigment layer of the eye is located in the zone of the ‘fibrous retina’ as well as between the proximal parts of the crystalline cones. There is a well developed proximal pigment layer in the very base of the retina, where also the proximal retinula cell nuclei are located.

Fig. 13 F. minima, sagittal section of compound eye cc: crystalline cone; co: corneal lenses; dp: distal pigment layer; fr: ‘fibrous retina’; pp: proximal pigment layer.

Remarks The Fraus eye structure in its main features conforms with that described previously from members of the Hepialidae s. str.; noteworthy traits like the elongate cones and the well developed proximal pigment layer are probably plesiomorphic at the lepidopteran level (Ehnbom 1948; Yagi and Koyama 1963; Kristensen 1968b). The three hepialid genera described by Ehnbom differ in details of retinula structure; Fraus is overall similar to Korscheltellus (Ehnbom’s Hepiolus lupulinus), while differing from Hepialus (Ehnbom’s Hepiolus humuli) in the non-aggregation of the retinula cell nuclei in the distal part of the ‘fibrous retinal layer’, and from Phymatopus (Ehnbom’s Hepiolus hecta) in the non-extension of the ‘rhabdom layer’ sensu Ehnbom (i.e. the non-tapered proximal retinula cell parts) towards the cones.

Antenna

Integument The antennae are relatively long, approximately 0.3-0.5 length of forewing in males, shorter in females. The scape is markedly wider than the remaining antennal segments and c. 1.5 times as wide as the pedicel, which is short, rounded and c. 1.5 times as wide as the flagellum segments. The intercalary sclerite (Figs 14, 15) is tongue-shaped, slender and almost parallel-sided, without microtrichia and lowered into the lumen of the scape. Flagellum with 39 (9. pteromela) to 55 (15. biloba) segments; several species have about 41 segments; females have one or two fewer segments than males. F. pilosa has short, cylindrical segments each with one broadly rounded lobe (Figs 25, 26). All other species have bipectinate antennae: the ‘shaft’ of each segment is long and cylindrical with rami arising from about midlength. The rami of 17. nanus and 18. furcata are relatively short and broad (Figs 23, 24, 35), while the rami of all other species are long and slender (Figs 21, 22, 27, 28, 33, 34, 36); in several species the rami are up to four times as long as the width of the ‘shaft’. Females of all species have markedly shorter rami than the males and the rami of both males and females are directed somewhat inwardly (Figs 21-28). The rami of the two rows are of subequal length or the anterior rami are longer than the posterior ones. The lateral/dorsal surface of the flagellomere shaft is covered with slender, elongate lamellar scales (Figs 7, 8, 16, 17); the vestiture is sparse or absent in some species, particularly in some species of the simulans group; the scales of the flagellum are longer and narrower in females (Fig. 17) than males (Fig. 16). The antennal scales have second order ridges and the scale sockets are scattered on the lateral/dorsal surface of the shaft.

The entire flagellum surface is covered with close-set fluted, tapering, curved cuticular projections; those on the medial surface are slender and microtrichia-like (Figs 19, 20, 31), while those on the dorsal surface of the rami are broader, almost scale-like with a pointed tip (Fig. 29).

Slender, curved sensilla trichodea are the dominant type of sensilla; they arise from the midline in the proximal part of the medial surface of the shaft (Figs 18, 19) and from the underside and lateral parts of the rami (Figs 18, 31); their length and density differs markedly between species (Figs 18, 33-36). There are few, one to four, sensilla auricillica on the distal side of the base of the rami on the medial surface of the shaft (Figs 19, 20). There are a few sensilla chaetica on the dorsal surface of the rami; 21. crocea has one before middle, a subapical and an apical sensillum chaeticum (Figs 29, 30); the number and position of the sensilla chaetica differ between species.

Musculature There are four bundles of extrinsic antennal muscles:

MM. H1-4, which all originate on the anterior tentorium (including on the dorsal arm), and insert on the base of the scapus.

Similarly, there are four groups of intrinsic muscles:

MM. H5-8, which all originate in the basal part of the scapus and insert on the base of the pedicellus.

Remarks Fraus is the only non-hepialid (s. str.) exoporian taxon with bipectinate antennae. Many genera of Hepialidae s. str. have bipectinate antennae (Quail 1900 - who also discussed and illustrated (p. 431, pl. 5, figs 8, 8a) the antenna of a Fraus species (as Hectomanes fusca)); almost all the genera with bipectinate antennae occur in the southern hemisphere and the only exclusive northern hemisphere genus with distinctly bipectinate antennae is Bipectilus Chu & Wang, 1985. Bipectinate antennae are not otherwise found among non-ditrysian Lepidoptera. Since the phylogeny of the hepialid (s. str.) genera is poorly understood it is difficult to assess whether bipectinate antennae is a hepialid (s. str.) groundplan trait.

Figs 14-20 Fraus, antennae - 14, 15: scape, intercalary sclerite (arrow) and pedicel of F. pteromela (14 - showing intercalary sclerite from broad surface) and F. simulans (15 - showing intercalary sclerite from side) - 16-20: flagellum of F. crocea at midlength; 16: lateral surface in male, 17: lateral surface in female, 18: medial surface in male, 19: sensilla and microtrichia on medial surface, 20: details of 19; st: sensillum trichodeum; sa: sensillum auricillicum.

Figs 21-28 Fraus species, cross-sections (schematic) of antennal flagellum at midlength - 21, 22: F. minima male (21), female (22) - 23, 24: F. nanus male (23), female (24) - 25, 26: F. pilosa male (25), female (26) - 27, 28: F. simulans male (27), female (28).

Figs 29-36 Fraus species, antennae - 29-32: F. crocea; 29: dorsal surface of ramus, 30: detail of 29, 31: ventral surface of ramus, 32: detail of 31 - 33: F. minima male, flagellum at midlength - 34: F. serrata male, flagellum at midlength - 35: F. nanus male, flagellum at midlength - 36: F. simulans male, flagellum at midlength; sc: sensillum chaeticum.

However, there is at present little evidence in support of this: several presumably primitive Hepialidae s. str. have filiform antennae. Nielsen and Robinson (1983: 46) showed that bipectinate antennae have developed twice within the genus Dalaca Walker, 1856 and that a single, subordinate species of Callipielus Butler, 1882 has bipectinate antennae. There are other examples among the Hepialidae s. str. of filiform and bipectinate antennae occuring in closely related taxa (Nielsen unpublished).

Flower and Helson (1976) first noticed that the flagellum surface in Hepialidae s. str. is covered with a dense mat of cuticular projections. This has now been observed in several hepialid (s. str.) genera and in Osrhoes Druce (Nielsen unpublished), Neotheora (Kristensen 1978a), Afrotheora (Nielsen and Scoble 1986) and Fraus. This surface structure may prove to be an additional hepialoid autapomorphy: Mnesarchaea has antennae with a network of fine ridges (Nielsen unpublished) as is typical for the Lepidoptera.

The presence of four tentorio-scapal, and four scapo-pedicellar muscles represents the glossatan groundplan condition.

Mandible

Integument The mandible is a small, but distinct, rounded lobe in minima (Figs 37-39). It is distinctly, though not strongly, sclerotized and bears a single, apparently basiconic sensillum. In polyspila the mandible is, relatively, markedly smaller and somewhat less sclerotized, but in some simulans specimens it is more strongly developed than in the minima examined.

Musculature (Fig. 54). M.H9. O: on the posterior cranial surface close to the inflection which sets off the postlabial ‘wart’. I: on a tubular apodeme which is invaginated behind the posteromedial margin of the mandible. The apodeme has a subbasal melanized zone. Note: M.H9 is presumably the mandibular abductor. Its antagonist, the cranial abductor, is completely reduced, but a small cuticular knob (with a central cavity) invaginated in the latero-anterior articular membrane of the mandible (Fig. 39) is presumably the vestige of its apodeme.

M.H10. O: on the lateroventral surface of the expanded part of the anterior tentorial arm. I: on the medial mandibular margin. Note: This very slender muscle (unequivocally observed in only a few of the sectioned specimens) is associated with a swelling of the mandibular nerve, presumably consisting of proprioceptive neurons.

Remarks Mandibular vestiges are widespread in the Hepialoidea, and they may be relatively larger in some members of the Hepialidae s. str. than in any Fraus species we have examined (Mortimer 1965; unpublished observations).

The single mandibular muscle hitherto known from adult Neolepidoptera has usually been considered a vestigial abductor (see, e.g. Denis and Bitsch 1973). We would a priori assume it to be unlikely that the much larger (in the lepidopteran ground plan, as in mandibulate insects generally) adductor should be first reduced as a consequence of mandibular reduction. The presence, in Fraus of an apparent vestige of the abductor apodeme laterad from the insertion of M.H9 supports this view. The origin site of M.H9 does not seem to be in conflict with this interpretation. Kristensen (1984f) showed that the tentorial mandible adductor is, after all, present in the lepidopteran ground plan, and the finding of this muscle, in an exceedingly reduced state, in a neolepidopteran moth is not particularly surprising.

Maxilla

Integument The maxilla in 1. minima (Figs 2, 4, 40, 41, 45-49) has a prominent basal piece (probably a cardo+ basistipes) which is an arched and well sclerotized plate. The tubular dististipes (i.e., the remainder of the maxillary base) bears a sclerotized plate proximally on the ventral surface and is otherwise softwalled. The palp consists of a single segment (Fig. 41), or, sometimes, what appears to be two indistinctly separated segments; it bears very small microtrichia and narrow lamellar scales. The galeal proboscis vestige is about 0.8x the head capsule length; it is held more or less curved, but we have seen no indication that it is capable of tight coiling. The inner galea wall (Fig. 47) has a non-spinose, transversely wrinkled surface which only near the base shows a faint longitudinal depression, i.e., a vestigial food groove. There are no sensilla on this inner surface except at the apex (Fig. 48). The outer wall is strongly spinose and bears scattered basiconic sensilla (Fig. 45). The dorsomedial spines constitute a functional linking mechanism (Fig. 46).

The basal maxillary piece is well developed in all species examined.

The proboscis vestige is prominent in all members of the ‘primitive grade’species groups (Figs 41, 45-49). Its size has been measured only in a few taxa besides minima; thus it is about 0.6x head capsule length in 15. biloba, and it equals head capsule length in 17. nanus. The palp may be distinctly two-segmented as in 8. orientalis (Philpott 1927b, as bilineata).

Figs 37-40 F. minima, mouthparts - 37: mandible and lateral part of maxilla - 38: sagittal section of mandible and maxilla - 39: horizontal (slightly oblique, dorsalmost level in left of figure) section of mouth region - 40: paramedial sagittal section of mouth region; ab: vestigial apodeme of mandibular abductor; bm: basal maxillary piece; ci: cibarial cavity; cm: cibarial muscularis; cp: cibarial plate; ga: galea; md: mandible; mp: maxillary palp: ph: prelabio-hypopharyngeal lobe; st: stipes. Muscle abbreviations in text.

Figs 41-44 Fraus, mouthparts - 41: F. minima, maxillary dististipes with palp and proboscis base - 42: F. crocea, as preceding; inset: proboscis vestige from other side - 43: F. simulans, mouth region in ventral view showing labral bulge and maxillary dististipes with palp and proboscis vestige; inset: palp apex - 44: F. polyspila, maxilla in oblique anterodorsal view; ds: dististipes; lb: labral bulge; mp: maxillary palp: pr: proboscis; pv: proboscis vestige.

In all members of the simulans group s. lat. the proboscis vestige is very small and hidden, in situ, by the prelabio-hypopharyngeal lobe; in none of the dissected species (19. pilosa, 21. crocea, 22. simulans, 23. polyspila) does it reach even as far as the tip of the maxillary palp. The proboscis vestige may bear microtrichial spines like the palp, and it has a complement of prominent apical setae and/or some shorter sensilla. The palp may appear distinctly two-segmented in such species as crocea (Fig. 42) and polyspila, but whether the proximal part really is derived from a formerly discrete segment probably cannot be ascertained. The apex of the palp bears a very prominent styloconic sensillum in simulans (Fig. 43), long pointed setae and shorter sensilla in polyspila (Fig. 44), and piliform scales in crocea (Fig. 42).

Musculature Of the rich lepidopteran groundplan complement of extrinsic maxillary muscles only two sets are retained, but considering the presumed non-function of the mouthparts of the adult moths their size is remarkable. The intrinsic musculature is also reduced, but the presence of intrinsic proboscis fibres is noteworthy.

M.H11. O: on the lower part of the expansion of the anterior tentorial arm. I: in the basal piece of the maxilla.

M.H12. O: on the posterior cranial surface adjacent to the upper/posterior corners of the postlabial ‘wart’. I: on the membranous area between the medioproximal corner of the dististipital sclerite and the upper lateroproximal corner of the prelabio-hypopharyngeal lobe, occasionally apparently just extending into the latter. Note: the membranous insertion area of M.H12 is here assigned to the maxillary territory since comparisons with other lower Glossata (see below) lead us to interpret this muscle as the cranial flexor laciniae.

M.H13. O: on the membranous medial and medioventral wall of the dististipes (Fig. 40, occasional fibres on the medial edge of the dististipital sclerite). I: in the apex of the galeal proboscis. The fibres run close to the medial wall of each galea (Fig. 49).

Remarks The diversity of maxillary structure in Fraus is perhaps the major source of concern about the monophyly of the genus as now delimited (see Phylogeny chapter). In some members of the ‘primitive-grade’species groups it is as complete as in any taxon in the superfamily Hepialoidea, whereas in the simulans groups s. lat. it is as vestigial as in several Hepialidae s. str. (described by, e.g., Philpott 1927b; Tindale 1941; Mortimer 1965).

The basal maxillary piece (which for comparative reasons is considered a composite cardo+ basistipes, see Kristensen and Nielsen 1979: 118) is reported to be conspicuously arched and more or less extensively melanized in many hepialoid moths, including members of the Hepialidae s. str. that have the apical parts of the maxilla strongly reduced; it is particularly prominent in Bipectilus (see Philpott 1927b; Mortimer 1965; Robinson 1977; Nielsen 1988). It is conceivable that a tentorial muscle (the H11 homologue) remains well developed in these cases, as it does in 23. polyspila.

The interpretation of M.H12 as the flexor laciniae rather than a cranio-hypopharyngeal or -prelabial muscle is based on the fact that its apparent homologue in Eriocraniidae (muscle 26, Kristensen 1968b) and Lophocoronidae (M.H11, Nielsen and Kristensen in prep.) has the insertion more distinctly separate from the base of the prelabio-hypopharyngeal lobe. Moreover, these moths have no cranio/tentorio-hypopharyngeal/prelabial muscles, and indeed such muscles were most probably lost in the groundplan of the Glossata.

The retention of prominent proboscis vestiges in the ‘primitive-grade’species groups of Fraus, as noted before, is unique within the Hepialidae s. lat. Moreover, the longest (relative to head capsule length) Fraus proboscis now known is longer than that of any other hepialoid so far recorded, whereas Kristensen (1978a) believed the Neotheora proboscis to be the longest. However, the proboscis vestige in Fraus deviates from that of Neotheora in having essentially non-concave medial surfaces, and it thereby deviates more from the ancestral stage. It also differs from that of Neotheora, Prototheora and Anomoses in having the median surfaces non-spinose; this, of course, casts serious doubt upon the validity of the medial galea spinosity as a hepialoid autapomorphy (see Kristensen 1978a: 284, in conjunction with Kristensen 1979). On the other hand it is a possibility that the obliteration of large, discrete plates in the medial galea wall in the Hepialoidea (Kristensen 1978a) is an autapomorphy of the superfamily; such plates are retained in the Mnesarchaeidae (Kristensen 1968a, 1979) in a configuration similar to that of the Eriocraniidae (Kristensen 1968b; Davis 1978) and hence probably of the glossatan groundplan.

The gap between proboscis development in the ‘primitive grade’species groups and the simulans group s. lat. is remarkable; no intermediate stages between the ‘distinct’ proboscis vestiges of