Phylogeny and host-plant relationships of the Australian Myrtaceae leafmining moth genus Pectinivalva (Lepidoptera, Nepticulidae), with new subgenera and species

Abstract The phylogeny of the mainly Australian nepticulid genus Pectinivalva Scoble, 1983 is investigated on the basis of morphology, and a division into three monophyletic subgenera is proposed on the basis of these results. These subgenera (Pectinivalva, Casanovula Hoare, subgen. n. and Menurella Hoare, subgen. n. ) are described and diagnosed, the described species of Pectinivalva are assigned to them, and representative new species are described in each: Pectinivalva (Pectinivalva) mystaconota Hoare, sp. n., Pectinivalva (Casanovula) brevipalpa Hoare, sp. n., Pectinivalva (Casanovula) minotaurus Hoare, sp. n., Pectinivalva (Menurella) scotodes Hoare, sp. n., Pectinivalva (Menurella) acmenae Hoare, sp. n., Pectinivalva (Menurella) xenadelpha Van Nieukerken & Hoare, sp. n., Pectinivalva (Menurella) quintiniae Hoare & Van Nieukerken, sp. n., and Pectinivalva (Menurella) tribulatrix Van Nieukerken & Hoare, sp. n. Pectinivalva (Menurella) quintiniae (from Quintinia verdonii, Paracryphiaceae) is the first known member of the genus with a host-plant not belonging to Myrtaceae. Pectinivalva (Menurella) xenadelpha from Mt Gunung Lumut, Kalimantan, Borneo, is the first pectinivalvine reported from outside Australia. Keys to the subgenera of Nepticulidae known from Australia, based on adults, male and female genitalia, and larvae, are presented. Host-plant relationships of Pectinivalva are discussed with relation to the phylogeny, and a list of known host-plants of Pectinivalva, including hosts of undescribed species, is presented. DNA barcodes are provided for most of the new and several unnamed species.


introduction
The subfamily Pectinivalvinae was described by Scoble (1983) for a group of Australian nepticulids retaining a pectinifer on the valva in the male genitalia. The structure of the pectinifer is similar to that in Opostegidae, the sister-group of Nepticulidae (cf. van Nieukerken 1986;Davis 1989). The subfamily remained monobasic, with Pectinivalva Scoble, 1983 as sole genus, until a second genus, Roscidotoga, was described by Hoare (2000a). Roscidotoga has the pectinifer absent (or possibly reduced to a thickening along the apex of the valva), but shares other apomorphies with Pectinivalva that support its inclusion in Pectinivalvinae (Hoare 2000a; van Nieukerken et al. 2011). The monophyly of this expanded Pectinivalvinae has been questioned by Puplesis and Robinson (1999), who provisionally follow the classification of Puplesis (1994), recognizing two subfamilies, Nepticulinae (including taxa formerly treated as Pectinivalvinae and Nepticulini) and Trifurculinae (equivalent to Trifurculini of van Nieukerken (1986)). However, Puplesis' (1994) 'cladogram' is not based on cladistic principles, since he frequently lists alternative states of the same binary character as apomorphies for sister-clades (e.g., his 'apomorphies'10-16 are merely the plesiomorphic states respectively of apomorphies 36-40, 48 and 42). Therefore this classification is rejected here, and we follow van Nieukerken (1986), Johansson et al. (1990) and Hoare (2000a) in our concepts of Pectinivalvinae and Nepticulinae. However, the current cladistic analysis is not designed to test the monophyly of Pectinivalvinae; this issue will be treated in a more comprehensive analysis of the genera and subgenera of Nepticulidae to be presented elsewhere. Ongoing molecular studies (Regier et al., van Nieukerken et al., publication pending) also confirm the monophyly of Pectinivalvinae, but not of Nepticulinae. Whereas these results may challenge the division of Nepticulidae in two subfamilies, the Pectinivalvinae will remain as a clade, and the rank of that clade is of little consequence for the present study. Scoble (1983) recognized two subgroups within Pectinivalva, based on the shape of the male valva (rounded or roughly triangular). Hoare et al. (1997) made additional observations on the distinction between the two groups, which they called the P. commoni and P. funeralis groups, and assigned all described species to one or the other. Because of the existence of undescribed species that showed character-states of both groups, these authors did not erect a formal subgeneric distinction, in order to avoid the possibility of naming a paraphyletic taxon. The present analysis, which includes several of the 'intermediate' species (i.e. Pectinivalva brevipalpa,P. minotaurus,P. 219 and P. 226), was designed to answer the question of whether Pectinivalva falls into definable monophyletic groups. In particular, is the P. commoni group, which retains stereo-microscope. Series of images for each specimen were subsequently montaged using Helicon Focus and Zerene Stacker software packages. Post-editing work, i.e. replacing backgrounds, 'removing' pins, and inserting scale bars was done in Photoshop.
Photographs of moths, leafmines (jointly described species) and all genitalia slides were taken by EJvN with a Zeiss AxioCam (HR or MR5) digital camera attached respectively to a Zeiss Stemi SV11 stereo-microscope and a Zeiss Axioskop H, using Carl Zeiss AxioVision software (version 4), for some photographs using the module "Extended focus". Manipulation of photographs, using Adobe Photoshop® was kept to a minimum: disturbing conspicuous shades, protruding parts of pins, dust and air bubbles in slides were removed or obscured, black backgrounds smoothened.
Drawings were prepared by RJBH, using a drawing tube. Drawings on one plate are not necessarily in the same scale.
Dna barcodes. DNA was extracted from caterpillars or from dry adult abdomens. DNA extraction from larvae was usually destructive; from abdomens and some larvae the non-destructive protocol by Knölke et al. (2005) was followed, allowing the preparation of the genitalia or larval skin as well. Details of methods are presented by van Nieukerken et al. (2012); we provide here the COI DNA barcode for several named and unnamed species of Pectinivalva, collected in 2000 and 2004 by the authors, and material on loan, collected by L. Kaila and co-workers. Barcodes of Roscidotoga were published before (van Nieukerken et al. 2011). Details can be found on the Barcode of Life webpages (http://www.barcodinglife.com/views/login.php) under the project "Nepticulidae -Pectinivalvinae Public Records [NEPPP]". Specimen data are also given in the online Appendix 3.
An NJ tree was prepared with Paup 4.0b10 for Windows (Swofford 2003), using uncorrected P distance (Srivathsan and Meier 2012). As outgroups we used Stigmella anomalella (Goeze, 1783) and Opostega salaciella (Treitschke, 1833) cladistic analysis. choice of terminal taxa: Species were chosen with two overall aims in mind: (a) to cover the morphological diversity of Pectinivalva as far as possible, and (b) to cover the host-plant range as fully as possible so that inferences could be made about the evolution of host-plant choice in the genus. Only species for which both sexes of the adults were available were chosen: this criterion excluded most of the previously described species. Again, for the most part, species were chosen for which preserved larvae and mines were available; however, the early stages of P. caenodora (Meyrick, 1906) are unknown. The early stages of P. commoni Scoble, 1983 have not been preserved, but a very similar species (P. 142: see Appendix 3) with genitalia very close to those of P. commoni was reared during the course of this study and the scores for the immature stages of P. commoni given here are based on the study of this second species.
Details of material examined for species not described in this paper are given in the online Appendix 3. Undescribed species which have been reared are referred to here by their generic name and a rearing number. One new species that has not been reared was also included: this is P. mystaconota sp. n. Pectinivalva mystaconota diverges in morphology from most typical Pectinivalva species in that the pectinifer on the male valva is replaced by rows of strong flattened setae (Fig. 40).
The results of an unpublished phylogenetic analysis including most subgenera of Nepticulidae (Hoare and van Nieukerken, in prep.) and molecular studies (van Nieukerken et al. in prep) suggest that Roscidotoga represents the sister-group of Pectinivalva (cf. Hoare 2000a). Roscidotoga callicomae Hoare, 2000 was therefore included as an outgroup, alongside a representative of Nepticulinae (Enteucha acetosae (Stainton, 1854)) and a representative of Opostegidae (Notiopostega atrata Davis, 1989). cladistic analysis. A maximum parsimony analysis with bootstrap was carried out with Paup 4.0b10 for Windows (Swofford 2003). For the heuristic search the branch swapping algorithm used was tree-bisection-reconnection (TBR). A bootstrap analysis was run for 200 replicates, each with 100 addition-sequence replicates.
Characters were traced onto the most parsimonious trees using the program Mesquite 2.7 (Maddison and Maddison 2009) to generate lists of apomorphies for monophyletic groups. Where the position of a character state change on the tree is ambiguous, the listed apomorphy is annotated according to whether it assumes accelerated transformation (ACC) or delayed transformation (DEL) of states and the alternative interpretations are given. The characters and their states are listed in Appendix 1.

Results and list of apomorphies
A heuristic search in PAUP produced 147 equally most parsimonious trees of length 60 with CI = 0.6667, RI = 0.8450 and RC = 0.5633. The strict consensus tree is presented in Fig. 1 (for bootstrap support see Fig. 2). Pectinivalva is consistently recovered as a monophyletic group with moderate support (bootstrap=76%), and within the genus, the three subgenera are almost always recovered, with strong support for Pectinivalva s.s. (92%) and Menurella (81%) (Fig. 1), but no support for Casanovula. Pectinivalva brevipalpa is the most problematic species, remaining in an unresolved polytomy with the three subgenera in the strict consensus tree. The 50% majority rule consensus tree is presented in Fig. 2, with bootstrap values; in this P. brevipalpa forms a clade with P. minotaurus, P. 219 and P. 226 in Casanovula. This placement is argued for below, and the 50% majority rule tree is used as the basis for the list of apomorphies given here, even without bootstrap support for Casanovula.
The monophyly of the genus Pectinivalva is supported by the following apomorphies:  Uncus with a pair of well-defined tufts of setae.  Vestibulum of female genitalia with a pair of lateral sclerites. The sclerites are absent in the clade P. mystaconota + P. 138 + P. 163; this is unambiguously reconstructed as a secondary loss.  Corpus bursae of female genitalia with extensive pectinations. This character is somewhat weak, since pectinations are present in many Nepticulinae (e.g. Stigmella and Acalyptris spp. ent and that the sculpturing of the cuticle in Pectinivalva s. str. is homologous to the spines of other Nepticulidae. However, there are two reasons for rejecting this idea. Firstly, the sculpturing is particularly marked on the prothorax, where other species of Pectinivalvinae lack any spines (although some Nepticulinae have a spiny larval prothorax). Secondly, the sculpturing is present on the prothoracic sternite, an area that is never spined in other nepticulids. The two characters are therefore considered to be independent.
The following possible apomorphies are problematic: (5-0) Forewing venation with R2+3 present. This is recovered as an unambiguous apomorphy of Pectinivalva s. str. in the analysis, since all other Pectinivalva species and all three outgroup taxa lack R2+3. However, loss of veins probably occurs rather easily in the evolution of these tiny moths, and independent losses are likely. Regaining a lost vein may not be impossible as supposed, e.g., by Meyrick (1898), but since R2+3 is present in most genera and subgenera of Nepticulinae except Enteucha, the character state evolution recovered here may be an artifice of taxon choice. (15-0 DEL) Cathrema of aedeagus supported by 2 or 3 interconnected sclerites. This form of cathrema is paralleled in P. brevipalpa, and therefore could equally be interpreted as an apomorphy of Pectinivalva as a whole, with subsequent reduction in P. minotaurus + P. 219 + P. 226 (ACC) and fusion into a sclerotised tube in Menurella.
The subgenera Menurella and Casanovula together form a monophyletic group supported by the following synapomorphies (but without bootstrap support): (9-1) Uncus bifid. The undivided, hood-like or V-or Y-shaped uncus of most Nepticulidae (including Pectinivalva s. str.) appears to be the plesiomorphic state in the family (cf. Scoble 1983;van Nieukerken 1986). (21-1) Larval antenna 2-segmented (second and third segments fused). The antenna in Pectinivalva s. str. and in Roscidotoga is 3-segmented, as it is in most Lepidoptera. The basal state of this character for Pectinivalva is recovered as ambiguous, due to the reduced larval antennae of two of the outgroups (Notiopostega with two segments and Enteucha with one). However, independent reductions in the antenna are considered far more likely than the regaining of a lost segment, so this is regarded as a robust synapomorphy of the two groups. In P. quintiniae, a further reduction to one segment has occurred, in parallel with Nepticulinae. (24-1) Prothoracic sternite of larva much longer than broad. The Opostegidae lack a ventral prothoracic sclerotization, but most Nepticulidae, including Roscidotoga and Pectinivalva s. str., have a rather broad sclerite in this position.
The monophyly of Casanovula has no bootstrap support. However, the last two of the three characters listed below are both unique within Pectinivalva and constant within this clade, so they are considered sufficient evidence to name this clade as a subgenus. The hostplant range is also distinctive (see below). Monophyly of the subgenus is supported by the following apomorphies: (1-1 ACC) Basal flagellar segments of male antenna expanded and flattened. The flattened antenna occurs in P. brevipalpa and P. minotaurus, but not in the other two species of the group included in the analysis. Since this antennal character is unique amongst Nepticulidae, it seems very unlikely to have evolved independently in P. brevipalpa and minotaurus, and is therefore best reconstructed as an apomorphy of Casanovula in the tree topology in Fig. 2, with subsequent loss in P. 219 + P. 226. The unresolved position of P. brevipalpa in the strict consensus tree ( Fig. 1) is due to the pyriform larval head , an apomorphy that it shares with Pectinivalva s. str. This is straightforwardly regarded as a parallelism, whereas the antennal character is not. For this reason, and because it shares the other two apomorphies listed below, P. brevipalpa is confidently assigned to Casanovula.  Forewing metallic, with shining fascia. Many Nepticulinae have a similar pattern, but as other species of Pectinivalva have drab, more or less unicolorous forewings, such coloration is apomorphic within the genus. One species of this group, not treated here (but included in the key to subgenera below) lacks a transverse fascia, but retains weakly metallic forewings with purplish reflections. Reflective coloration is known to correspond in Lepidoptera with diurnal activity, and members of this group have not been collected at light, in contrast with members of Pectinivalva s. str. and Menurella. The small eyes of the known species (interocular index less than 0.7) further confirm that the moths are diurnal.  Corpus bursae of female genitalia lacking signum. All other species of Pectinivalva have a well developed signum. Although a signum is also lacking in many opostegids, including Notiopostega, and in the nepticulid outgroups Roscidotoga and Enteucha, the presence of a very similar form of signum (the 'toothed band') in Pectinivalva s. str. and in three plesiomorphic species of Menurella (P. acmenae, xenadelpha and quintiniae) strongly suggests that the loss of the structure in Casanovula is apomorphic.
The monophyly of Menurella has strong bootstrap support (81%). Members of the group share the following apomorphies:  Pectinifer of male valva with fewer than 20 elements. There are more than 20 elements in the pectinifer in all species of Pectinivalva s. str. and Casanovula that retain the structure, and usually more than 20 in the Opostegidae (e.g. 45 in Notiopostega atrata, 35 in Eosopostega issikii Davis, 1989). The pectinifer elements have probably been lost more than once in Menurella : they are absent, e.g., from P. quintiniae and P. warburtonensis (Wilson). Conversely, several species of Pectinivalva in ANIC (not treated here) clearly belong to Menurella on the basis of apomorphies 15-1, 16-2, 17-1 and 18-2 (see below), but have a pectinifer with more than 20 elements.
(15-1) Cathrema of aedeagus supported by a smooth sclerotized tube. The tube is always present within the group and not found in other pectinivalvines. It could possibly represent a fusion of the sclerites associated with the cathrema in Pectinivalva s. str. A similar smooth tubular structure is associated with the cathrema in species of Enteucha (see van Nieukerken 1986), although not in E. acetosae; this is presumed to be a parallelism.  Lateral sclerites of female vestibulum strongly developed. These sclerites are always narrow (occasionally absent) in Pectinivalva s. str. and Casanovula. In P. acmenae they are still relatively narrow, but have deeply forked tips. In the remaining species of Menurella the sclerites are very broad and robust.  Pectinations of corpus bursae of female restricted to posterior part of corpus.
More extensive pectinations are found in most other species of Pectinivalva s. str. and Casanovula (except P. 138 and P. 163).
Within the three subgenera, some species-groups are recovered in all most parsimonious trees. Within Pectinivalva s. str., P. caenodora and P. 34 form a monophyletic unit with 98% bootstrap support (here referred to as the P. caenodora group). At least one other species (P. 89), not included in the current analysis, is known from this group. P. caenodora and P. 34 share the following apomorphies:  Forewing with a pale costal streak. This streak is lacking in the third member of the group mentioned above. (10-2) Tufts of setae on dorsum of uncus mounted on lobes. These lobes are found in other species of Pectinivalva s. str. which were not included in the current analysis and do not share the other synapomorpies of P. caenodora + P. 34: hence this apomorphy may define a broader species-group which includes the P. caenodora group.  Central element of gnathos broad and cordate. Shared by the third member of the P. caenodora group, but again also present in a few other species of the P. commoni group. (14-1) Sublateral processes of transtilla strongly reduced.  Signum continuously toothed.
Pectinivalva 163, P. 138 and P. mystaconota (bootstrap 78%) share the following apomorphies: (6-1) Male hindwing strongly expanded at base. This is paralleled in Menurella in the group of species including P. scotodes, P. 2 and P. 119. (7-1) Male hindwing with androconial pocket. Again, this structure is present in the Menurella species listed above and in P. 91. (16-0) Vestibulum of female lacking lateral sclerites. The sclerites, which constitute an autapomorphy of Pectinivalva, are most parsimoniously regarded as secondarily lost in these species.
The position of P. commoni and P. 5 within Pectinivalva s. str. cannot be resolved on the basis of the characters used in the current analysis. P. commoni most closely resembles the group formed by P. 163 + P. 138 + P. mystaconota on the basis of head and forewing colour. It also shares with these species the presence of androconial scales on the male hindwing (Hoare et al. 1997: fig. 5), although it does not have a true androconial pocket. Pectinivalva 5 most closely resembles the members of the P. caenodora group on the basis of the narrow valva with a strongly expanded apex (cf. Hoare et al. 1997: fig. 38): however, this could represent a plesiomorphic condition.
A further problematic synapomorphy is: (15-2 ACC) Cathrema moderately developed, with 0-1 associated sclerites. Since this state is shared with Roscidotoga, the presumed sister-group of Pectinivalva, it may alternatively be a retained plesiomorphy (DEL) (see above, under Pectinivalva). In the latter case, Pectinivalva s. str.and P. brevipalpa have independently converged on a weaker cathrema with 2-3 sclerites.
Pectinivalva 219 and P. 226 (76% bootstrap support) share the following behavioural apomorphy: (28-1) Exit hole a large slit; larva pupating in mine. This habit has not been observed in any other species of Pectinivalvinae.
Pectinivalva 226 has only one pair of D setae on the larval mesothorax ; the state of this character is unknown in P. 219, which was reared from pupae. It may prove to be a further apomorphy of this subgroup (paralleled in Pectinivalva s. str. and in P. 91).
Within Menurella, Pectinivalva acmenae and P. quintiniae form an unresolved basal trichotomy with a clade containing the remaining species. The remaining species in the current analysis (P. scotodes, P. 2, P. 91 and P. 119; 85% bootstrap support) share the following apomorphies: (7-1) Male hindwing with androconial pocket. This character is paralleled in P. mystaconota + P. 138 + P. 163. (13-1 ACC) Pectinifer elements of male valva broad and tooth-like. The pectinifer elements are narrow  in P. 2, which is interpreted as a reversal. Alternatively, broad elements must have evolved independently in P. 91 and in P. scotodes + P. 119.  Signum in form of parallel spinules. This form of signum is highly characteristic of this group and not known elsewhere in Nepticulidae.
Pectinivalva tribulatrix (not included in the analysis) shares all these characters and also belongs to this group.
Amongst the described species of Pectinivalva, P. funeralis (Meyrick, 1906) and P. libera (Meyrick, 1906) (each known only from the male) share apomorphy 4-1, and resemble P. scotodes and P. 119 in their male genitalia; they are therefore considered to belong to the same species group within Menurella.

Phylogeny and classification of Pectinivalvinae: conclusions
The genus Pectinivalva falls into three monophyletic clades, each described and diagnosed as a subgenus of Pectinivalva below: subgenus Pectinivalva Scoble, 1983, subgenus Casanovula subgen. n. and subgenus Menurella subgen. n. It would be possible to regard these as informal species-groups. However, a more profound subdivision seems necessary, because all three subgenera contain recognizable species-groups (some of them outlined above), which are likely to be of use in the future, especially in the highly speciose subgenera Pectinivalva and Menurella. Two other possible classifications would have been consistent with the results of the cladistic analysis: (i) To give all three subgenera generic rank; (ii) to give Pectinivalva (sensu stricto) generic rank, and to erect a new genus containing Casanovula and Menurella as subgenera. Genera and subgenera are essentially artificial and subjective concepts: therefore the only criteria for judging between such classifications are considerations such as usefulness, informativeness and consistency of ranking within the family (i.e. a subgenus in Pectinivalva should be roughly equivalent to a subgenus in Ectoedemia, although sister-groups need not have the same taxonomic rank). Admittedly, the use of subgenera can be rather cumbersome; however, it also allows one to convey more information about the relationships of a species in its name. Because Scoble (1983) and van Nieukerken (1986) are followed here and elsewhere (Hoare 2000b) in recognizing subgenera in the Trifurculini, and because the degree of distinction between the three groups within Pectinivalva as here defined is comparable to that between trifurculine subgenera, the most conservative classification has been adopted.
The hypothesis that Roscidotoga falls outside Pectinivalva, and is not derived from within that genus, is corroborated by the cladistic analysis.

host-plant relationships of Pectinivalva
The host-plant relationships of Roscidotoga were discussed by Hoare (2000a) and van Nieukerken et al. (2011); their hosts are in the Cunoniaceae and Elaeocarpaceae, which belong to the Oxalidales in the eurosid I or fabid clade of the eudicots (Stevens 2008;APG III 2009). The eurosid I clade was of key significance in the early evolution of phytophagous insects (Ward et al. 2003), including Lepidoptera, and is the most important plant group in the host spectrum of Nepticulidae (Menken et al. 2010). Cunoniaceae and Elaeocarpaceae can both be considered ancient Gondwanan families based on current distribution and diversity (see discussion and references in Hoare 2000a).
As stated before, almost all known host-plants of Pectinivalva belong to the Myrtaceae, but one species is known from Paracryphiaceae (Quintinia) (see below for discussion). A full list of known host-plants is provided in Appendix 2; this includes myrtaceous hosts from which no moths have been reared, but where nepticulid mines assumed to belong to Pectinivalva have been collected. The assumption is based on the fact that no other genera of Nepticulidae have been reared from Myrtaceae in Australia. The Myrtaceae is an old Gondwanan family, with a long history in the southern continents (Johnson and Briggs 1981). Myrtaceae belongs to the order Myrtales in the eurosid II or malvid clade of eudicots (Stevens 2008;APG III 2009). The family reaches its greatest diversity, both at the generic and specific levels, in Australia: Eucalyptus L'Hérit., with approximately 800 species, dominates the vegetation over much of the continent. The dominance of Myrtaceae is certainly a relatively recent phenomenon, associated with the drying of the Australian climate during the mid to late Tertiary, and the replacement of rainforest with sclerophyllous vegetation (cf. Carpenter et al. 1994;Christophel 1994). Nevertheless, the association of Pectinivalva s.l. with myrtaceous hosts may predate this diversification considerably. A fossil myrtaceous leaf from the early Oligocene (ca. 35 million years B.P.) of Cethana, Tasmania (Carpenter et al. 1994: fig. 12.3e) shows an unmistakeable nepticulid mine, very similar, for example, to that of P. (Casanovula) brevipalpa.
The discovery of one species of Pectinivalva (quintiniae) feeding on a host other than Myrtaceae is an interesting development. P. quintiniae is in the subgenus Menurella, and although it apparently occupies a basal position in this subgenus (Fig. 2), it remains most parsimonious to assume that its host relationship is the result of a secondary shift from Myrtaceae. P. quintiniae is a rainforest species. Quintinia verdonii F. Muell., the host-plant of P. quintiniae, has been recently referred to the family Paracryphiaceae (APG III 2009). Winkworth et al. (2008) recovered Paracryphiaceae as sister-group to the Dipsacales, and the family now has its own order, Paracryphiales in the campanulids clade within the core eudicots (Stevens 2008;APG III 2009). Three genera are placed in the family: Quintinia DC., with 25 species distributed in New Guinea, Australia, New Caledonia and the Philippines, Sphenostemon Baill., with ten species occurring in New Guinea, Australia and New Caledonia, and Paracryphia Baker f., which is monotypic and endemic to New Caledonia. The distribution is con-sistent with a Gondwanan origin for the family, although there exists a late Cretaceous fossil from Sweden (Silvianthemum suecicum Friis) that has features in common with Quintinia (Friis 1990). Related groups, according to the phylogeny of Winkworth et al. (2008) (i.e. Dipsacales, Apiales), contain few known host-plants of Nepticulidae (van Nieukerken 1986).
Each subgenus of Pectinivalva has a distinctive host-plant range. Subgenus Pectinivalva has only been reared from species of Eucalyptus, the most speciose genus of Myrtaceae in Australia. Casanovula is the only subgenus without species on Eucalyptus: and Syncarpia Ten. in Syncarpieae cannot yet been associated with one of the subgenera, although we predict that at least the mines on Myrteae most likely belong to Menurella.
In general the observation that closely-related moths tend to feed on closely-related plants (cf. van Nieukerken 1986;Hoare 2000a) is corroborated. This is clearly the case for subgenus Pectinivalva, where all species apparently feed on a single genus of hosts. It also applies to the P. (C.) 219 species group of Casanovula, all of whose species feed on Melaleuca; the inclusion of Callistemon R. Br. in this genus by Craven (2006) is followed here. Pectinivalva (C.) brevipalpa and P. (C.) minotaurus appear very similar and closely related, although not recovered as sister-species in the cladistic analysis; however, whilst their host-plants (respectively Tristaniopsis and Lophostemon) were formerly placed together in the genus Tristania R. Br., recent molecular phylogenetic evidence from the matK gene suggests that they belong in different tribes of Myrtaceae (Kanieae and Lophostemoneae) (Wilson et al. 2005;Biffin et al. 2010). Finally, within subgenus Menurella, the group of relatively 'derived' species (P. (M.) 91 + P. (M.) scotodes Hoare + P. (M.) 119), which have a pectinifer consisting of broad tooth-like elements, all feed on Eucalyptus and closely-related myrtaceous genera in the tribe Eucalypteae of Wilson et al. (2005), with the exception of P. (M.) tribulatrix, which feeds on Rhodomyrtus in the tribe Myrteae. Angophora, the host genus of P. (M.) 119, is sister to Corymbia (Hill and Johnson 1995;Wilson et al. 2005); at least one of the two (unnamed) species of Pectinivalva to have been reared from Corymbia also belongs within this 'derived' group of subgenus Menurella on the basis of the pectinifer.
The phylogeny suggests that the common ancestor of modern Pectinivalva was a Myrtaceae-feeder, and that the presence of a species on Quintinia is the result of a host-shift. It seems most likely that the split between Roscidotoga and Pectinivalva predates the Miocene aridification of Australia (ca. 24 -ca. 5 million years B.P.), and that the original hosts of Pectinivalva (like those of Roscidotoga) were rainforest plants. It is interesting that rainforest Myrtaceae still host species of subgenera Casanovula (brevipalpa and minotaurus) and Menurella (acmenae, xenadelpha, quintiniae and tribulatrix); the position of these species in the phylogeny is consistent with the hypothesis that their hosts may be ecologically and/or phylogenetically close to the original Pectinivalva host-plant.
The subgenus Pectinivalva appears to lack representatives on rainforest plants, and has only been reared from the sclerophyllous genus Eucalyptus. The earliest macrofossils currently accepted as belonging to Eucalyptus are leaves of E. kitsoni Deane from the Berwick Quarry, Victoria; these date from the late Oligocene or very early Miocene (ca. 25 million years B.P.), a time when south-east Australia was probably beginning to dry climatically, though still dominated by rainforest (Pole et al. 1993). However, the clade to which Eucalyptus belongs (tribe Eucalypteae of Wilson et al. 2005) may have originated as long ago as the Cretaceous, as evidenced by the presence of the relictual genus Arillastrum Panch. ex Baillon in New Caledonia (Ladiges et al. 2003), although the relaxed molecular clock analysis by Biffin et al. (2010) gives a range of dates from late Cretaceous to Eocene for the split of Eucalypteae from Syncarpieae. A date for the split between Pectinivalva s.s. and the other subgenera (e.g. from molecular data) would be of great interest to shed further light on the pattern of host-plant choice.

Dna barcodes
We provide the DNA barcodes ( Fig. 3) for seven of the eight new species named in this paper (only missing P. brevipalpa), and for one previously named species (P. caenodora). In addition we include barcodes for a further nine or ten unnamed species that were available for barcoding. The main reason for including barcodes is to aid recognition of species and identification of immature stages without rearing. We do not intend to use these data for phylogenetic purposes, for which data on more genes are required.

Subfamily Pectinivalvinae Scoble
Description. Scoble (1983) gave a detailed description of the Pectinivalvinae, but recently much more material of this subfamily (especially immature stages) has become available, and the concept of the Pectinivalvinae has been expanded by Hoare (2000a) to include the previously unrecognized genus Roscidotoga. For these reasons a complete redescription of the subfamily is provided here. A revised diagnosis of the subfamilies of Nepticulidae is provided by Hoare (2000a: table 1).
Adults. Head : Labial palpi 2-or 3-segmented; galeae short; maxillary palpi 5-segmented; antennae with sensillum vesiculocladum usually or always 5-branched (needs more detailed study in some species). Collar usually consisting of piliform scales. Forewing: underside sometimes with androconial scales in male; subdorsal retinaculum absent. Hindwing: upperside often with androconial scales in male. Wing venation (Figs 31-36): forewing without closed cell, Cu present, long; 1+2A unthickened, running obliquely from base of wing to meet dorsum before tornus; hindwing with trunk of Rs+M usually more or less deflected towards costa. Abdomen sometimes with specialized scales dorsally in male; S2a more or less pentagonal, usually with transverse rows of minute spines. Legs: fore-tibia of males sometimes thickened with specialized scales.
Cocoon. Usually reddish brown; usually spun outside the mine. Pupa. Head: Clypeus squarish; frons with a pair of conspicuous setae posteriorly; labial palpi distinctly longer than maxillae. Eclosion more or less dorsal, so that suture between eyecaps and frons remains largely intact ventrally. Abdominal segments 2-8 each with 3-4 rows of spines on dorsum, and a prominent pair of dorsal setae.
Distribution. Australia, Borneo. Probably more widespread in Australian and Oriental regions than currently known. anazona (Meyrick, 1906) funeralis (Meyrick, 1906) libera (Meyrick, 1906) planetis (Meyrick, 1906) primigena (Meyrick, 1906) trepida (Meyrick, 1906) warburtonensis (Wilson, 1939)  Type species. Pectinivalva commoni Scoble, 1983, by original designation. A large and diverse genus, here subdivided into three subgenera on the basis of the phylogenetic reconstruction presented above. The following overview of the morphology of the genus should be taken in conjunction with the more detailed descriptions of the subgenera given below. Because of the great number of species in the genus, a complete revision is impractical at present. Species have been selected for description in order to represent the range of host-plants, morphology and distribution so far known in Pectinivalva.

checklist of Pectinivalvinae
Description. Adults. Head capsule : labial palpi 2-or 3-segmented. Underside of forewing and upperside of hindwing often with androconial scales in male. Costal bristles of male hindwing absent or replaced by lamellate scales. Legs: foretibia of male sometimes thickened with specialized scales. Upperside of abdomen sometimes with androconial scales in male. Anterior edge of T2 weakly sclerotized medially.
Female genitalia (Figs 73-103). S8 usually very broad and squared off. Vestibulum usually with a pair of lateral sclerites associated with apophyses anteriores. Corpus bursae well sclerotized, without diverticulum, usually with single signum.
Pupa. As described for subfamily. biology. Most known larvae leaf-miners on Myrtaceae, with one species on Paracryphiaceae (Quintinia A. DC.).

Diagnosis.
Distinguished from Roscidotoga, the only other known genus of Pectinivalvinae, externally by the forewing pattern (without silver streak from mid-costa or suffusion of metallic scales towards apex); in the male genitalia by the presence of a gnathos and a well-sclerotized uncus with strong tufts of setae; and in the female genitalia by the simple (unexpanded) apophyses anteriores and the well-sclerotized corpus bursae, which lacks a diverticulum.
Distribution. Australia (known from all states and territories), Borneo (a single species, P. xenadelpha, described below).
Diagnosis. See Table 1. Distribution. Australia (known from all states and territories). included species. In addition to the six previously and newly described species, also approximately 65 undescribed species in the anic, of which the following, cited by their anic rearing numbers, have been studied in detail for the current work: P. Discussion. Pectinivalva (Pectinivalva) is a relatively diverse subgenus, and could probably be subdivided into several species groups. One such group, the P. (P.) caenodora group, was diagnosed above. We do not propose to erect any further named species groups here, but we describe below a species that diverges strongly from most other members of the subgenus, and has several close relatives. Their placement in P. (Pectinivalva) is argued for below, but as the larvae are as yet unknown, this decision may have to be revised.
Diagnosis. Both sexes can be distinguished from similar members of the subgenus by the black eyecaps with their weakly scaled transparent bases. In addition, the combination of the pointed valvae with their long, strong setae (visible in undissected  , emg. 11.x.1995. males) and the moustache-like patch of hair-scales on T5 of the abdomen is characteristic of the male. In the female genitalia, the absence of lateral sclerites in the vestibulum and the transversely rugose corpus bursae are diagnostic.
Distribution. Collected in scattered localities in eastern Australia from Wellington, N.S.W. south to Mt Nelson, Hobart, Tasmania; presumably widespread, but not yet known from Victoria.
Dna barcode. RMNH.INS.24106, Genbank KC292479 Derivation. The specific name is derived from the Greek mystax (a moustache) and notos (a back) and refers to the tuft of hair-scales on T5 in the male. It is an adjective.
Remarks. Several species related to P. (P.) mystaconota are known: all lack a pectinifer and have more or less dense tufts of setae on the dorsal surface of the valva. The group seems to be best represented in Western Australia. As the larvae are unknown, and the only definite apomorphies for the subgenus Pectinivalva are characters of the larva, the assignment of this group to the subgenus remains to be confirmed. The un- divided uncus and the form of the sclerites associated with the cathrema in the male genitalia, and the form of the signum in the female genitalia, are characteristic of the subgenus Pectinivalva, but these features may be plesiomorphic within the genus as a whole. However, in the most parsimonious trees resulting from the cladistic analysis presented above, P. mystaconota was placed as sister species to P. (P.) 138 + P. (P.) 163. For these reasons, it is here placed in the subgenus Pectinivalva.
biology. Host-plants: Lophostemon Peter G. Wilson spp., Tristaniopsis Peter G. Wilson spp., and Melaleuca L. spp. (including species formerly assigned to Callistemon R. Br.) (all Myrtaceae). Mine (Figs 117, 118): either a narrow gallery more or less filled with frass, or a gallery expanding into a blotch; exit-hole a small semicircular slit, a small semicircular hole, or a large slit: in the last case larva pupating in mine.

Distribution.
Known only from eastern Australia: Queensland, N.S.W., A.C.T. and Tasmania; to be expected in Victoria.
Derivation. The subgenus is named (in the diminutive) after the famous Italian adventurer and philanderer Giacomo Casanova, in reference to the unusual sexual ornamentation of the males of some species (e.g. P. (C.) minotaurus sp. n., in which the male has strongly dilated antennae and specialized scales on the abdomen upperside). It is considered feminine (in spite of its derivation) to accord with the gender of Pectinivalva.
included species. No previously described species are referable to this subgenus. The following species are described below: Pectinivalva (C.) brevipalpa sp. n. and P. (C.) minotaurus sp. n. Also at least five undescribed species in the anic, of which the Discussion. This is the least speciose of the three subgenera of Pectinivalva. Two species-groups can conveniently be recognised. In the P. (C.) brevipalpa group (not monophyletic according to the cladistic analysis), the antenna of the male is dilated and flattened at the base, the vertex bears a pair of sclerotized crests, and the labial palpus is modified, with segments 2 and 3 reduced or fused. Abdominal sternite 2a is strongly spinose. The host-plants are Tristaniopsis and Lophostemon spp. and pupation is outside the mine. In addition to P. brevipalpa and P. minotaurus, a single undescribed species (P. 41, feeding on Lophostemon confertus), is referable to this species group. In the P. (C.) 219 group, the labial palpi and the vertex are unmodified; S2a lacks spines; the host-plants belong to Melaleuca (including Callistemon), and pupation may be within the mine. About four species are known in this species group. Description. Male (Fig. 6). Wingspan 4.3-5.9 mm. Head capsule (Figs 23-26): labial palpi reduced, 2-segmented; maxillary palpi with ratio of segments from base approximately 0.2: 0.4: 0.5: 1.2: 1.0; interocular index 0.67; vertex with a pair of sclerotized crests. Frontal tuft ferruginous; collar white; eyecaps white; antennae with basal segments dilated and flattened, gradually tapering, shining lead-grey, whitish beneath, ca. 32 segments. Thorax and tegulae dark fuscous with purplish reflections. Forewing to 2/3 dark fuscous with purplish reflections; a shining silver to pale golden fascia at 2/3, slightly broader on costa, apex of wing dark fuscous without reflections; cilia pale grey beyond a line of fuscous-tipped scales. Hindwing grey, unmodified; cilia grey. Abdomen lead-grey, slightly shining. Female (Fig. 7). Wingspan 4.3-5.2 mm. Similar to male, but antennae not dilated at base, 18 segments.
biology. Host plant: Tristaniopsis collina Peter G. Wilson & Waterhouse (Myrtaceae). Egg: on underside of leaf. Mine (Fig. 117): commences as very long narrow gallery either filled with greenish frass or with black linear frass, broadens rather abruptly into gallery with central line of black frass; exit-hole on upperside, a semicircular slit. Cocoon: reddish brown. Occupied mines have been collected on 25 June, 1 July, 13 July and 3 August.
Diagnosis. The male is superficially similar to that of P. (C.) minotaurus, but differs in its much less strongly expanded antennae. The male of P. (C.) minotaurus also differs in having shell-like androconial scales on the upperside of the abdomen, visible on dissection, and a more distinctly H-shaped vinculum (Figs 46, 62). The female of P. brevipalpa is also very similiar to that of minotaurus but can be distinguished on dissection by the presence of the indistinct sclerites ½ way down the corpus bursae.
Distribution. New South Wales. Derivation. The specific name is derived from the Latin brevis (short) and palpus (the sensitive palm of the hand: hence, in zoology, a palp) and refers to the reduced, 2-segmented labial palpi of the adult male. It is an adjective.  8.iv.1906, [A.J. Turner], slide 11506 (anic).
Larva. Green. Head (Fig. 106) (Fig. 118): commences as very long narrow gallery with black linear frass, leaving narrow clear margins, broadens rather abruptly into an irregular wide gallery or elongate blotch, sometimes with gallery parts, with central line of black frass or in the case of the blotch, frass concentrated on one or both sides; exit-hole on underside, an almost circular hole. Cocoon (Fig. 125) Derivation. The species is named after the famous beast of Greek mythology, the Minotaur. The name (a noun in apposition) refers to the extraordinarily expanded and flattened male antennae, which are likened to the Minotaur's horns.
Remarks. The antennae of the male are the most strongly modified of any known species of Nepticulidae. Although many male-specific head structures in other insects are utilized in male-male competitive interactions over mates (e.g. the lateral cephalic projections of Phytalmia spp, Tephritidae (Moulds 1977)), such direct competition is unknown in Lepidoptera, and the antennae of minotaurus are more likely to function in close-range courtship, along with the androconial scales on the male abdomen. Similar widened flagellomeres are known from the genus Thisizima Walker, 1864 in Tineidae (Yang et al. 2012). The androconial scales are also remarkable, two distinct types being present in contiguous patches on the abdominal dorsum.
Description. Adults. Head capsule (Figs 21,22): labial palpi 3-segmented; interocular index 0.49-0.84. Antennae in male occasionally broadened in middle, or with pedicel and segment 1 of flagellum modified as in Fig. 22. Collar consisting of piliform scales (lamellate scales in one undescribed species). Wingspan 3.2-7.0 mm. Thorax and forewing usually unicolorous greyish to fuscous, or with transverse pale fascia or opposite pale spots on costa and tornus at 2/3, occasionally yellowish with dark markings. Costa of forewing in male sometimes with a tuft of very narrow stiff scales towards base. Hindwing of male occasionally expanded at base; androconial pocket often present. Underside of forewing in male sometimes with androconia. Wing venation (Figs 34-36): R2+3 in forewing absent. Abdomen: S2a with or without spines. Legs: fore-tibia in male of those species with androconial pocket usually thickened above with blackish scales.
Female genitalia ( (Figs 119-124): usually a narrow gallery more or less filled with frass, occasionally a very short gallery leading to and enveloped by a blotch; exit-hole usually a small semicircular hole.
Derivation. The subgeneric name is the diminutive of Menura, the genus to which the lyre-bird belongs. It stems from a fancied resemblance between the uncus in some species of the group and the tail of the male lyre-bird. It should be treated as feminine.
included species. In addition to eleven described and new species, also approximately 70 undescribed species in the anic, of which the following, cited by their anic rearing numbers, have been studied in detail for the current work: P.   acmenae, P. xenadelpha, P. tribulatrix, and P. quintiniae with unusual morphology and host-plant.
biology. Host plants: Eucalyptus pilularis Smith, E. carnea R. Baker, E. acmenoides Schauer and probably E. saligna Smith (see below) (Myrtaceae). Egg: on upperside of leaf. Mine (Fig. 119): commences as a tight spiral around the egg, causing a raised red-brown spot on the leaf about 3-5 mm in diameter; later broadens into a more or less contorted linear gallery with black frass leaving narrow clear margins; exit-hole on leaf underside, a crescentic hole. Often several mines to a leaf. Cocoon: reddish brown. Occupied mines were collected on 17 and20 July 1995, and21 Dec 1996, and have also been recorded in January, March, April, June and August (Moore 1966). Derivation. The specific name (an adjective) is derived from the Greek skotodes, meaning either 'dark' or 'dizzy'. It refers both to the blackish coloration of the adult male moth, and to the habit of the young larva, which mines in tight circles.
Remarks. The two female paratypes reared from mines on Eucalyptus carnea collected near Brisbane have a sparser scattering of brown scales on the yellow ground colour than the females reared from E. pilularis. However, no other differences have been observed between specimens from these two host-plants, and the mines also appear to be identical.
This species was first collected by K.M. Moore, who described and illustrated the mine (Moore 1966: figs 15, 15A). There are specimens in the anic (here designated paratypes) reared by him in the 1950's. The host-plant is indicated on the labels only by a rearing number; mines from his herbarium with the corresponding number are all in leaves of Eucalyptus acmenoides. He also recorded mines on E. saligna, but no specimens reared from this host-plant have been located. Moore referred to this species as 'Nepticula sp. 3' and regarded it as related to N. gilva Meyrick. He was probably misled by the wing-pattern of the female of P. (M.) scotodes, which bears some resemblance to that of P. (P.) gilva: he would not have seen the type specimen of P. (P.) gilva in the BMNH. The two species are not closely related and belong to different subgenera of Pectinivalva.  10-21.x.1995;2♂, 36.19S, 150.03E, Mt Dromedary, N.S.W., emg. 22, 24.x.1995, R.J.B. Hoare, E.S. Nielsen and M.J. Matthews, genitalia slides 10213, 11242 (anic);2♂, 28.42S,153.37E, Broken Head NR, N.S.W., 13.vii.2000, emg. 15-18.viii.2000.
Description. Male (Fig. 12).  mm. Head: frontal tuft ferruginous; collar inconspicuous, consisting of white, grey-tipped scales; eyecaps anteriorly white, posteriorly shining grey with bluish reflections; antennae shining dark grey, whitish beneath, ca. 35 segments. Thorax, tegulae and forewing uniform shining dark grey with strong blue reflections; an inconspicuous tornal spot consisting of a few white scales; cilia dark grey. Hindwing unmodified, pale grey; cilia pale grey. Under- side: forewing grey with faint brassy reflections; hindwing grey. Abdomen shining dark grey; anal tuft inconspicuous, dark grey. Female (Fig. 13).  mm. Similar to male, but antenna with 23-25 segments, and forewing rather broader. Wing venation as in Fig. 36. Abdominal tip not as broad and 'square' as in females of other Pectinivalva spp.
Male genitalia (Figs 52-54, 66). Capsule ca. 425 μm long, forming a narrow triangle. Anterior edge of vinculum excavated in a half-oblong. Tegumen rounded, with ventral extensions on each side overlapping lateral arms of gnathos. Uncus rectangular, bilobed, lobes slightly produced, with 3 setae on each. Gnathos central element long, reaching just beyond uncus, ending in small swelling. Valva (Fig. 53) ca. 210 μm long, squarish, more rounded caudally and produced into a short point at exterior corner of apex; apical ½ with numerous spine-like setae on dorsal surface; pectinifer consisting of ca. 18 spine-like elements. Long sublateral processes present. Juxta a weak subcircular plate. Aedeagus (Figs 54,66) ca. 510 μm long, a curved spine arising towards apex on left, a shorter spine to right of this one, a third spine in line with second and anterior to it. Vesica basally with cathrema surrounded by a field of many broad, short cornuti; a separate field of ca. 9 long narrow cornuti above opening of ejaculatory duct.
Larva. Green. Length of head ca. 440 μm; width ca. 350 μm. Thorax: prothoracic sternite in shape of Y with expanded base (Fig. 113); an additional small roundish sclerite on each side of this and antero-dorsal to SV and V group of setae. Chaetotaxy and spinosity: T2 with 11 pairs of setae (L3 present); otherwise as described for subgenus Casanovula. Anal rods distinctly forked posteriorly.
Diagnosis. This is the one of three known species of Pectinivalva in which the forewings have a bluish lustre but no transverse fascia. The others are P. xenadelpha and P. quintiniae, both described and diagnosed below. There is an undescribed Australian species of Stigmella which sometimes occurs together with P. (M.) acmenae, and in which the forewings are similarly unicolorous dark blue; however, the Stigmella species is distinctly larger (wingspan 6-8 mm), and has a collar consisting of white lamellate scales; its larva is a leaf-miner on Baloghia inophylla (G. Forster) P. Green (Euphorbiaceae).
Distribution. New South Wales. Vacated mines probably of this species were seen abundantly along the coast near Manley, Sydney.
Dna barcode. RMNH.INS.23541, Genbank KC292474. Derivation. The specific name is derived from the former host-plant genus, and is a noun in the genitive. Because the moth is referred to under this manuscript name in the first author's unpublished thesis, we have chosen to retain it for consistency, in spite of the change in classification of the host-plant.
Female genitalia (Figs 98-100). Total length ca. 770 μm. T9 prominent, with a group of 5 setae on each side. Apophyses anteriores rather narrow, curved inwards; apophyses posteriores narrow, straight, longer than anteriores. Lateral sclerotizations of vestibulum strongly developed, but not forked. Ductus spermathecae with ca 6 close set convolutions. Posterior part of corpus normal, slightly folded, without markings; anterior part rounded, with rows of inconspicuous pectinations; signum consisting of broken linear sclerotization surrounded by oval sclerotized ring with blunt dentitions.
Larva. Not preserved. biology. Host-plant: Syzygium acuminatissimum (Blume) A. DC. (Myrtaceae) (formerly Acmena acuminatissima), very closely related to the Australian S. smithii (Biffin et al. 2006), a widespread species in the mountains of south Asia, from India and China to New Guinea and the Pacific islands (Chen and Craven 2007). Egg: almost invariably on upperside, almost always on or near midrib. Mine (Fig. 123): a long, very narrow contorted gallery, first half very narrow, running from midrib to leaf margin, or sometimes along midrib, filled with blackish frass, second half much wider, much contorted, often zigzagging, frass compact, black, leaving narrow clear margins; exit-hole on underside, a semicircular hole. Cocoon: ochreous. Occupied mines were collected on 18 and 20 November; they occurred together with abundant mines of a Heliozela species.
Diagnosis. Very similar externally to P. (M.) acmenae from Australia, but lacks the pale tornal spot of that species. In the genitalia, the lateral sclerites of the vestibulum are not forked (as they are in acmenae), and the apophyses posteriores are distinctly longer than the apophyses anteriores (same length in acmenae).
Distribution. Borneo, East Kalimantan: Gunung Lumut. Derivation. The species name (a noun in apposition) derives from the Greek xenos (stranger, foreigner) and adelpha (sister) and refers to the close relationship to acmenae as well as the great geographical distance between this and other known Pectinivalva species.
Remarks. We choose to describe this species here, even on the basis of a single female, to be able to record the genus from outside Australia. The detailed knowledge of its life history and three DNA markers (including CO1 barcode) will make future association with males straightforward.  56) ca. 305-320 μm long, rounded caudally; apically fringed with numerous spine-like setae on dorsal surface; pectinifer absent, but apex of valva thickened and well-sclerotized. Long sublateral processes present. Juxta a subrectangular plate. Aedeagus (Figs 57, 69) ca. 455-480 μm long, a curved spine arising towards apex on left. Vesica basally with many broad, short cornuti, grading into field of much larger longer cornuti towards apex.
Larva. Green. Head as in Fig. 108 Derivation. The specific name (a noun in the genitive) is derived from the hostplant genus.
Female genitalia (Fig. 101-103). Total length ca. 335 μm. T9 produced on each side into prominent anal papillae, each with a group of 7 setae. Apophyses anteriores moderately narrow, curved inwards; apophyses posteriores narrow, straight, longer than anteriores. Lateral sclerotizations of vestibulum strongly developed, forked, the bifurcations diverging widely. Ductus spermathecae with 6 convolutions. Corpus small, about as long as wide, folded, covered with many pectinations; signum of concentric bands of fence-like spinules, indistinct.  Larva. Green. Fieldnotes state that it feeds with dorsum upwards, which may be incorrect. Larva not preserved.
biology. Host-plant: Rhodomyrtus macrocarpa Benth., finger cherry (Myrtaceae). Many mines and three larvae were collected on the ca. 20 cm long leaves of seedling shrubs. Egg: on either side of leaf. Mine (Fig. 124): a narrow, long gallery, either completely meandering, or partly straight and following a major vein; frass black, broken and dispersed over total gallery width, not leaving clear margins; edges of gallery not straight, irregular; exit-hole on underside, a semicircular to oval hole. Cocoon reddish brown. Occupied mines have been collected on 22 July.

Diagnosis.
One of the smallest Pectinivalva species we know, recognised by un-  5 No females of Trifurcula have yet been captured in Australia, and the distinction used here is based on Holarctic members of the genus (see Johansson et al. 1990).  7 6 No constant morphological differences have been found between larvae of P. (Casanovula) and those of P. (Menurella). 7 No larva of Trifurcula has been found in Australia, and the distinction used here is based on the description of European species in Johansson et al. (1990).