A new species of Paralucia Waterhouse & Turner, 1905 (Lepidoptera: Lycaenidae) from the highlands of south‐eastern Australia

A new species of lycaenid butterfly, Paralucia crosbyi sp. nov. (Theclinae: Luciini), is described, diagnosed and illustrated from Namadgi National Park, ACT, and adjacent areas in New South Wales in the highlands of south‐eastern Australia. It is most similar to Paralucia spinifera E.D. Edwards & Common, 1978, but comparative morphology of the adult and immature stages indicates fundamental differences between the two species. Maximum Likelihood phylogenetic analysis of mitochondrial cytochrome oxidase subunit I (COI) of all Paralucia species recovered P. crosbyi sp. nov. and P. spinifera as reciprocally monophyletic, with a mean uncorrected ‘p’ pairwise divergence of 0.93%. Paralucia crosbyi sp. nov. appears to be a narrow‐range endemic, restricted to dry montane eucalypt open woodland or woodland between 920 and 1130 m asl, in which the mean annual rainfall varies from 700 to 800 mm and where an abundance of the larval food plant Bursaria spinosa Cav. subsp. lasiophylla (E.M.Benn.) L. Cayzer, Crisp & I. Telford (Pittosporaceae) and colonies of the attendant ant Anonychomyrma sp. (itinerans species group) (Dolichoderinae) are established. The immature stages are described, illustrated and compared with those of P. spinifera and P. aurifera (Blanchard, [1848]). Despite being limited to high altitudes, adults fly in late winter–early spring (late July to early October). The species is univoltine, with much of the year (~9 months) passed in the pupal stage. The ecology, biology and life cycle are discussed, and a likely mode of speciation is hypothesised.


INTRODUCTION
The thecline genus Paralucia Waterhouse & Turner, 1905 is a small lineage endemic to Australia (Braby 2000;Edwards 1996;Edwards et al. 2001).It is closely related to Pseudodipsas C. & R. Felder, 1860, Lucia Swainson, 1833 and Acrodipsas Sands, 1980, which Eliot (1973) placed in the Lucia section of the tribe Luciini.Eliot diagnosed this group of butterflies according to three-character states: male fore tarsus pointed, male genitalia with a 2-pronged uncus and male genitalia with a prominent Y-shaped juxta.In Paralucia, the hind wing has the termen without tails, but the tornus is strongly produced, especially in the male, and the male is without androconia (sex scales).
In 2021, a new population of Paralucia, provisionally identified as P. spinifera (Bond & Vardon 2022), was discovered in the mountains of the Australian Capital Territory (ACT), approximately 250 km SSW of the known sites of P. spinifera in the central highlands (Bathurst-Lithgow-Oberon area).The population was first detected in Namadgi National Park, ACT, along the Brandy Flat Fire Trail (1030 m) at 1142 h AEST on 12 August 2021 by Susan Wishart.A series of photographs of a male and a female were uploaded to the Canberra Nature Map (2021) and provisionally identified as Zizina otis (Fabricius, 1787) on 16 August 2021.The site was recovering from the January 2020 megafire storm that swept through Namadgi NP, and the vegetation was in an early stage of ecological succession.The following day, moderator Suzi Bond forwarded the images to me for clarification, suspecting that the individuals may be lycaenids other than Z. otis.Examination of the photographs of the female revealed several features (bronze and purple colouration on dorsal forewing, black bar at end of forewing discal cell, rounded hindwing without tornal projection, series of blue subterminal markings on both wings and orangebrown tipped antenna), which indicated that it was a Paralucia, most likely P. spinifera or a species closely allied to P. spinifera.Subsequently, I visited the area in early September 2021 and succeeded in collecting a series of adults and the immature stages approximately 1 km from the original discovery site.Additional material was subsequently collected from two other sites in the park.A closer examination and comparison of this material with P. spinifera have revealed many significant differences.Hence, the purpose of this paper is to describe, diagnose and illustrate this new taxon.

Molecular data
Specimens prepared for sequencing were collected as live adults from the field and preserved in the Australian National Insect Collection, Canberra (ANIC).Each specimen was assigned a unique voucher number: The body was preserved in 100% ethanol stored at À20 C, while the wings were detached and preserved in paper envelopes in the pinned collection.One or two legs were removed and sequenced for each specimen.Mitogenomic DNA was extracted and sequenced from nine specimens according to the method outlined in Braby et al. (2023).The mitochondrial gene cytochrome oxidase subunit I (COI) (1500 bp) was then selected from each of these samples.Sequence data were assembled and aligned as amino acid sequences in AliView (ver. 1.28; see http://ormbunkar.se/aliview;Larsson 2014).The dataset was then combined with publicly available sequence data on NCBI GenBank (https://www.ncbi.nlm.nih.gov/genbank/) for the 'barcode' fragment of COI (658 bp) for 12 specimens of Paralucia spp.and four species of Luciini (Lucia limbaria, Acrodipsas myrmecophila, Pseudodipsas eone and Pseudodipsas cephenes), which were used as outgroups.Thus, the final data matrix comprised 25 sequences representing four ingroup taxa and four outgroup taxa (Table 1).All new sequence data are available from NCBI GenBank.

Morphological data
Material of the adult and immature stages of Paralucia spp. was collected from Lithgow, NSW, in 1996 and2022;Tallaganda National Park (E of Hoskinstown), NSW, in 2004and 2017-2022;and Namadgi National Park, ACT (three sites: Brandy Flat FT;Rendezvous Ck;and Nursery Swamp Tk) during several visits between 2021 and 2023.Adults were collected using a standard entomological sweep net or reared from the immature stages.Eggs, larvae and/or pupae were collected from all sites.Eggs were also obtained in captivity by enclosing a mated female in a small plastic container with a cutting of the larval food plant, Bursaria spinosa, that had been in contact with the attendant ant, Anonychomyrma sp.(itinerans species group).Larvae were reared in captivity with or without the attendant ants inside plastic containers supplied with fresh cuttings of the larval food plant, which were changed every 1-2 weeks.
The male genitalia were studied and compared by dissecting two specimens of each taxon as follows: Paralucia crosbyi sp.'Wahronga, NSW, 12 Nov. 1954, D. Sands' (ANIC).Dissection of the genitalia was performed using standard techniques: The entire abdomen was removed, placed in 10% KOH, boiled for 20 min and then transferred to 30% ethanol for cleaning, dissection and examination.Completed dissections were fixed in 95% ethanol and then 100% isopropanol.The genitalia were placed in an alcohol-based gel in an excavated glass block, photographed using a Leica M205A microscope and stacked using Helicon Focus 5.3 according to the technique of Su (2016).Terminology for genitalia follows Klots (1970) and Eliot (1973).
Adult size was determined by measuring the length of the forewing, that is, the straight-line distance between the apex and base (i.e., the point of attachment of the wing with the thorax), using a digital calliper to an accuracy of 0.1 mm.
All material examined was in the ANIC.

RESULTS
The ML tree based on mitochondrial COI recovered Paralucia as monophyletic with strong to moderate support (Figure 1).The monophyly of each species of Paralucia

Diagnosis
Paralucia crosbyi (Figure 2a-h) is distinguished from P. spinifera (Figure 3a-h) by the following nine-character states: (1) The dorsal iridescent ground colour of males is bright bluish purple or violet in P. crosbyi, whereas it is pinkish purple or bronze-purple with more conspicuous black margins in P. spinifera.(2) In both sexes, the discal bar (which comprises a double row of black lines) on the dorsal forewing is generally broader and more conspicuous in P. crosbyi than in P. spinifera.(3) Females of P. crosbyi are monomorphic with only a single colour morph (purple), whereas females of P. spinifera are polymorphic with two colour morphs: brown (Figure 3e) and purple (Figure 3g).The brown morph of P. spinifera may have a few purple scales in the basal and/or subbasal areas of both wings.Moreover, in P. spinifera, the predominate morph is brown, while the purple morph occurs in low frequency (25%, n = 63).Among the female purple morphs, (4) the dorsal iridescent area is bluish purple or violet and generally extensive on both wings in P. crosbyi, whereas it is darker purple and, although variable in extent, generally reduced, especially on the hindwing, in P. spinifera; and (5) the dorsal ground colour, though variable, is generally paler brown in P. crosbyi, whereas it is richer dark brown in P. spinifera.(6) The series of pale blue subterminal spots between veins R 5 and 1A + 2A on the dorsal forewing are nearly always present in P. crosbyi females (94%, n = 35), whereas only about half (52%, n = 25) of the females of P. spinifera examined have the spots present-differences in these proportions are highly significant (χ 2 = 12.32, d.f.= 1, p < 0.001; Yates' correction for continuity).( 7) The dorsal hindwing of female P. crosbyi has a set of three distinct black subterminal 'teeth-like' triangular-shaped patches between veins M 3 and 1A + 2A proximal to the series of pale blue subterminal streaks, but these markings are greatly reduced in extent in P. spinifera.(8) In both sexes, the ventral ground colour in P. crosbyi is grey or grey-brown, whereas it is generally rich brown and rarely grey-brown in P. spinifera.(9) The ventral hindwing markings form a complex pattern in both species, but they are more clearly defined in P. spinifera, and the postmedian band is extensively suffused proximally and distally with orange or rusty brown, a feature that is greatly reduced in extent or absent in P. crosbyi, the difference being more pronounced in males.

Male
Head.Black with grey piliform scales; frons white with dark grey and black piliform scales; labial palp black dorsally, second and third segments white ventrally with black piliform scales, third segment black ventrally; eye black; antenna 7-8 mm long, flagellum with 31-34 segments (shaft 19-21, club 12-14), shaft black ringed with white, club black, tipped orange-brown, white towards base.
Thorax.Black dorsally with bluish grey hairs, white ventrally; legs grey-brown, fore tibia with long terminal spinelike process, mid and hind tibiae each with a pair of spines (tibial spurs), mid tibia shorter than first tarsal segment.
Forewing.Length 10.4-12.4mm (x = 10.80 mm ± 0.500 SD, n = 25); costa strongly arched near base, then straight; termen slightly rounded particularly towards apex; upperside colour iridescent blue-purple with broad black costal and terminal margins, black bar at end of discal cell; veins black; a row of up to six small pale blue subterminal spots between veins R 5 and 1A + 2A; cilia black, tipped white between veins; underside ground colour generally grey or occasionally grey-brown, slightly iridescent, with a series of darker bands edged black, then white; small subbasal band in discal cell; broken submedian band of two spots from R 1 to lA + 2A; broken median band of three spots from R 1 to CuA 2 , with section between R 2 and M 3 displaced distally towards end of discal cell; postmedian band from R 2 to lA + 2A, with section between M 1 and M 3 displaced distally; crenulate black subterminal line and dark brown terminal line enclosing a series of six indistinct dark brown marks between R 5 and 1A + 2A; cilia grey-brown, tipped white between veins.
Hindwing.Termen rounded; tornus slightly produced; upperside colour iridescent blue-purple with broad black costal and inner margins, narrower terminal margin; veins black; a row of five pale blue subterminal spots or streaks between veins M 1 and 1A + 2A; cilia black, tipped white between veins; base and dorsum with numerous long pale grey hairs; underside ground colour generally grey or occasionally grey-brown, slightly iridescent, often weakly suffused with rusty brown proximally and distally to postmedian band, and a complex series of darker bands and markings edged with black, then white; illdefined subbasal band from Sc + R 1 to 1A + 2A; broken submedian band of two or three spots from Sc + R 1 to 1A + 2A; broken median band of two spots from M 1 to M 3 at end of discal cell and from CuA 1 to 1A + 2A; postmedian band from Sc + R 1 to 3A, with section between Sc + R 1 and M 1 displaced proximally, section between M 1 and CuA 2 displaced distally; crenulate black subterminal line and dark brown terminal line enclosing a series of seven variable dark brown marks between Sc + R 1 and 1A + 2A; cilia grey-brown, tipped white between veins.
Forewing.Length 11.2-13.4mm (x = 11.86 mm ± 0.524 SD, n = 25); termen evenly rounded; upperside colour brown or pale brown, with basal and central area including discal cell and dorsum iridescent purple; prominent double black bar at end of discal cell; veins dark brown; a row of up to six pale blue subterminal spots or streaks between veins R 5 and 1A + 2A; black terminal line; cilia black, tipped white between veins; underside ground colour generally grey-brown or occasionally grey, slightly iridescent, markings similar to male; cilia grey-brown, tipped white between veins.
Hindwing.Upperside colour brown or pale brown, with basal and central area including discal cell iridescent purple; black bar at end of discal cell; a row of five pale blue subterminal spots or streaks between M 1 and 1A + 2A, the last three spots between veins M 3 and 1A + 2A each edged proximally with a black triangular-shaped mark; black terminal line; cilia black, tipped white between veins; base and dorsum with numerous long pale grey hairs; underside ground colour generally grey-brown or occasionally grey, slightly iridescent, markings similar to male; cilia grey-brown, tipped white between veins.Variation.The underside pattern is extremely variable in both sexes, especially the shading or colour of the dark bands and the extent to which they are edged with white (Figure 2b-d,f,h).In males, the upperside iridescent area varies from brilliant bluish purple to purple.In females, the extent of the dorsal iridescent purple area was more extensive in specimens reared in captivity that emerged in winter, whereas it was reduced in a single specimen that was reared in summer, suggesting that the extent of purple is influenced by environmental factors (e.g., temperature).In one wild-collected female specimen in early August, the iridescent colour is pale blue instead of purple.In females, the dorsal ground colour varies slightly from light brown to brown (Figure 2e,g).The number and size of the pale blue subterminal spots on the forewing vary in both sexes.In males, most specimens (86%, n = 36) have the forewing spots present, but the number varies from 1 to 6, although most specimens (77%) have 4 or 5 spots; occasionally the spots are reduced to just a few scales.In females, all but two specimens examined (n = 35) have the forewing spots present, but the number varies from 1 to 6, although most specimens (77%) have 4 or 5 spots, similar to males.

Immature stages
Egg (Figure 5a-c) Diameter 0.8 mm, height 0.4 mm; hemispherical, lime-green; micropyle deeply depressed; chorion with reticulated pattern of pits separated by ridges that bear conspicuous knobs or small projections at their junctions.posterior end of each segment, brown dorsolateral line, more pronounced and darker on AS7, followed by reddish lateral line and indistinct reddish ventrolateral line; thoracic and abdominal segments with colourless lateral setae, dorsal and lateral surfaces densely covered with numerous short pale cream secondary setae; abdominal segments 4-6 each with short brown and black dorsal setae at posterior end of segment; abdominal segment 7 with Newcomer's organ; abdominal segment 8 grey-green, with prominent dorsolateral projection bearing eversible organ, often everted laterally; abdominal segments 9-10 grey-green, with cream markings; anal plate mottled dull grey-black; spiracles black.

Pupa
(Figure 8a-c) Length 10 mm, width 4 mm; head and thorax brown; antennae distinctly blackened distally towards club; mesothorax with pair of darker brown subdorsal spots; wings pale yellowish brown, with numerous scattered brown dots; abdomen pale yellowish-brown, with darker brown or grey-brown middorsal line on abdominal segments 2-6, and scattered dark brown markings dorsolaterally; spiracles pale brown; surface of cuticle roughened with minute raised dots, particularly on head and mesothorax; a few anal hooks present, central girdle absent.

Remarks
Edwards and Common (1978) described Paralucia spinifera from Yetholme 16-20 km E of Bathurst, NSW (850-1000 m asl) based on six type specimens-the holotype male and five paratypes (2♂, 3♀).They illustrated the holotype male and a paratype female, both of which I have examined in the ANIC.
A characteristic feature of P. spinifera is the presence of a conspicuous, non-articulated spine-like process at the tip of the fore tibia of both sexes (Edwards & Common 1978), a morphological feature that is absent in both P. aurifera and P. pyrodiscus.However, this feature is also present in both sexes of P. crosbyi (Figure 4b,c).
In addition to the phenotypic differences in wing pattern elements noted above (see Diagnosis section), there are slight differences in the morphology of the immature stages.In particular, the egg of P. crosbyi (Figure 5a-c) is dark lime-green, whereas in P. spinifera (Figure 5g-i), it is pale bluish green.The egg of P. aurifera (Figure 5j-l) is of similar shape and size to P. crosbyi and P. spinifera, but the pits are smaller, the projections at the pit junctions are longer and spine-like and the colour is dull greenish yellow, later changing to grey.The chorion of P. crosbyi changes with age, with older eggs becoming white.Moreover, older eggs have a smoother surface, with the projections reduced or absent (Figure 5d-f), possibly due to exposure to the harsh climatic elements (i.e., weathering from ice/frost and intermittent snowfall).The first instar larvae of P. crosbyi (Figure 6a-d) and P. spinifera (Figure 6e) are identical, and they are very similar to P. aurifera (Figure 6f).The main difference lies in the number and length of the posterior setae on abdominal segment 10: In P. aurifera, there is a set of six posterior setae, four of which are longer and of approximately equal width, whereas in P. crosbyi and P. spinifera, there are only four setae, with the two inner setae considerably smaller (about half length) than the two outer setae.The final instar larvae of the three species have the same general colour pattern and markings, but in both P. crosbyi (Figure 7a-c) and P. spinifera (Figure 7d-f), the markings are richer and more brightly contrasted than in P. aurifera (Figure 7g-l), particularly the reddish oblique subdorsal and lateral lines; in P. aurifera, the colour pattern is paler with duller reddish lines.There are no apparent differences in the final instar larva between P. crosbyi and P. spinifera, and in both species, the extent of the black markings on abdominal segments 7 and 8 is variable.
Pupae of the two species are very similar, but in P. crosbyi (Figure 8a-c), the abdomen has a single dark brown or grey brown middorsal line, whereas in P. spinifera (Figure 8d-i), there is a pinkish brown middorsal line, which later changes to grey, often followed by one or two faint pink dorsolateral lines.The pupa of P. crosbyi has a series of brown dorsolateral markings on the abdomen, but these markings are reduced or absent in P. spinifera.The pupa of P. spinifera is more variable in colour, and it may also be pale greenish brown.The pupa of P. aurifera (Figure 8j-l) is of similar shape but is immediately distinguished by the colour pattern, particularly the browner coloured wings and conspicuous pinkish red longitudinal markings on the abdomen.In addition to the anal hooks, the pupa of P. aurifera is also attached by a central girdle, which is absent in the two other species.
No discernible differences in the male genitalia were found between P. crosbyi and P. spinifera.A comparison of P. spinifera and P. aurifera revealed that the male genitalia of these species are also very similar, but P. aurifera can be distinguished by the narrower tegumen and straighter posterior margin of the uncus (in lateral view)-these differences are clearly shown in the illustrations in Edwards and Common (1978, figs 9 and 11).The illustration of the male genitalia of P. pyrodiscus by Eliot (1973, fig. 27) is broadly similar to the three other species.

Etymology
The specific name crosbyi is an patronym of David F. Crosby, named in honour of his life-time contribution to the study of Victorian butterflies, including his pioneering work on the taxonomy and conservation of Paralucia (Crosby 1951(Crosby , 1987)).

Common name
The 'Violet Copper' is proposed as a common name, reflecting the brilliant deep bluish purple or violet dorsal colouration of the male.

Distribution
The species appears to be a narrow-range endemic, restricted to Namadgi National Park, ACT, and adjacent areas in New South Wales.In the ACT, it is currently known from >50 sites (S.Bond, pers. comm. 2023) at altitudes between 920 and 1130 m asl.In New South Wales, it has been recorded from Captains Flat (930 m) and the Tindery Mountains near Michelago east of the ACT border (S.Bond, pers. comm. 2023).Searches in similar environments further east in Tallaganda National Park, NSW (>1000 m) on the Great Dividing Range did not detect the species, despite an abundance of Bursaria spinosa and Anonychomyrma sp.(itinerans species group) ants-only adults and immatures of P. aurifera were found.Nielsen et al. (2023) gave the distribution as the 'Brindabella Range, ACT', but this location is erroneous; the species is known only from the mountain ranges further east in the lower rainfall zone.

Habitat and ecology
The breeding habitat consisted of montane eucalypt open woodland or woodland (>900 m asl) in the dry rainfall zone (mean annual rainfall 700-800 mm).In the Orroral Valley at Nursery Swamp Track, the habitat (Figure 9a,b) consisted of an open woodland with scattered trees of Eucalyptus viminalis growing on orogenic granite on a north-facing aspect above an extensive grassy frost hollow; the understorey comprised a mix of open rock pavement, grassland and shrubland dominated by the larval food plant Bursaria spinosa Cav.subsp.lasiophylla (E.M.Benn.)L. Cayzer, Crisp & I. Telford (Pittosporaceae) (Figure 9e), with some Banksia marginata and Acacia dealbata.Much of the habitat at this site was not burned by the January 2020 megafire storm.However, at Brandy Flat Fire Trail, the habitat (Figure 9c,d) consisted of a woodland dominated by Eucalyptus rubida, E. pauciflora, E. dives and E. bridgesiana on Late Ordovician undifferentiated sediments (greywacke, sandstone and slate) on a steep south-west-facing aspect; the larval food plant grew in abundance in a more floristically diverse shrubby understorey.The habitat was similar at Rendezvous Creek, except the woodland was dominated by E. rubida growing on granite.At all three sites, the larvae were obligatorily attended by the attendant ant, Anonychomyrma sp.(itinerans species group) (Dolichoderinae) (Figure 9f), which were relatively abundant.The site at Brandy Flat FT had been severely impacted by the this subspecies probably represents a clinal form of the nominotypical subspecies that it replaces on heavier clay soils at higher altitudes in the tablelands and lower mountain ranges in south-eastern Australia.
Thus, the four key components of the critical habitat appear to be the following: (1) high altitude (920-1130 m asl); (2) relatively low mean annual rainfall (700-800 mm); (3) co-occurrence of Bursaria spinosa ssp.lasiophylla and Anonychomyrma sp.(itinerans species group), typically in high density or abundance; and (4) vegetation structure with an open canopy and understorey (i.e., open woodland) allowing penetration of sunlight to the ground layer during morning and afternoon to facilitate basking and thermoregulation by adults (see below).
At all three sites studied, Paralucia aurifera was found to breed sympatrically with P. crosbyi, to be attended by the same species of ant and to utilise the same individual food plants.However, adults of P. aurifera flew later in the season (from late September or early October to late February or early March) so that the flight periods of the two species were temporally separated, with only a limited period of overlap in late September and early October when the abundance of P. crosbyi had greatly diminished.Larvae of the two species, however, occurred at the same time (September to February) and even occurred together on the same food plants.

Biology
Eggs were laid singly or very occasionally in pairs.They were generally laid on the ground near the base of the larval food plant (within a radius of 360 mm) on a variety of substrates, including stones, the underside of rocks, leaf litter (e.g., pieces of charcoal, sticks, dead eucalypt leaves and dry grass blades) and the leaves of herbs; however, they were occasionally laid on the underside of green leaves of the larval food plant.On hatching, the first instar larva chewed a hole at the top of the chorion, through which it escaped, but it did not consume the remainder of the eggshell.It is not clear how newly hatched larvae from eggs laid on the ground locate the foliage of the larval food plant.First and second instar larvae grazed the leaf surface (epidermis), whereas later instars chewed the entire leaf.The larvae (after instar II) fed nocturnally and were attended by large numbers of the attendant ants.When not feeding, they retreated to the base of the main stem/trunk of the larval food plant, usually just below ground level, where the attendant ant constructed a byre, or under small stones or dead branches close to the ant-byre.As the larvae grew, the ant-byre became enlarged, and sometimes it extended above the ground; during the day, the larvae remained hidden inside the ant-byre.On one occasion, during overcast conditions, a few mid-instar larvae were observed during the late afternoon (1700 h AEDT) foraging on the larval food plant attended by ants, but most larvae were at rest on the basal stem of the plant concealed inside the ant-byre.Food plants with high numbers of larvae became partially or even completely overgrazed (temporarily defoliated) (Figure 9e).The larvae pupated in ant galleries deep in the soil, usually within a 200 mm radius of the main stem/trunk of the larval food plant; they typically lay loose in the soil.Pupae were noted to stridulate, making a series of audible clicks, in captivity.
Adults (Figure 10a-f) flew only during sunshine, regardless of ambient temperature; when overcast, they settled on the ground, rocks or low objects with wings closed (Figure 10e,f).Both sexes regularly employed dorsal basking (Figure 10a-d) to thermoregulate by opening their wings directly towards the sun.When basking, the dorsal surface of the wings refracted extraordinary colours of bronze and iridescent green that are not visible in set museum specimens, especially males (Figure 10a) and, to a lesser extent, females (Figure 10d).Adults that emerged in captivity sat motionless with wings closed and did not move or fly until they were exposed to direct sunlight.Because adults are active when temperatures are relatively cold (i.e., during winter-early spring at high altitudes), access to sunlight to thermoregulate cannot be overstated.Males were active during the morning and afternoon, typically from around 1000-1500 h depending on access to sunlight and temperature; very few males were observed outside these times, but some individuals were active as late as 1600 h.They engaged in aerial contests and established mating territories in the breeding habitat by perching in open sunny patches on the ground or low down on herbs, shrubs, including the larval food plant, and other objects.A pair was observed mating at midday (1200 h AEST, 2 September 2021).Females resided close to the ground and were observed ovipositing only during the early afternoon (between 1305 and 1415 h).Before laying, females basked for some time to thermoregulate by settling on the ground or low down on the larval food plant.Mated females rejected courting males by flying in a slow downward 'hover' flight, rapidly beating their wings.Adults were rarely observed to feed on flowers-on one occasion, a female was observed feeding on the flowers of Hovea linearis (Fabaceae), and on another occasion, a male was observed feeding on the flowers of Leucopogon attenuatus (Ericaceae), both in mid-August.
Paralucia crosbyi is univoltine (Figure 11), with the broad flight period lasting approximately 11 weeks, from late July (earliest record: 2♂ 25 July 2023, S. Wishart, pers.comm.) to early October (latest record: 5♀ 9 October 2021).However, they were most abundant during August and September, and only a few adults were observed in July and October.Eggs were recorded from August to October, and larvae from September to February.Most larvae pupated by December or January, but a few larvae were found in the field as late as February.The duration of the egg stage in August was approximately 3 weeks (20 days, n = 6) under laboratory conditions; however, this is likely to be longer under field conditions.In captivity, larvae completed development in approximately 2.5 months, or 78 days on average (Table 3).The pupae underwent diapause, which lasted for up to 9 months.However, from a sample of larvae reared in captivity (n = 25), one pupa developed directly with a female emerging on 31 December 2021 after a pupal duration of approximately 2-3 weeks.It is not known if direct development occurs in the field, but no adults were recorded in December or January.

DISCUSSION
The discovery of Paralucia crosbyi in the ACT is remarkable given the proximity of its known distribution to Canberra, Australia's capital city, and underscores the valuable contribution made by citizen scientists.Paralucia crosbyi was completely unknown until it was discovered in 2021, when it was first detected (photographed) in August of that year by a citizen scientist (S.Wishart).The restricted montane distribution, with adults flying for only a limited period so early in the season (late winter-early spring), has no doubt contributed to it being overlooked.However, its presence following the megafire storms cannot be understated-the high abundance of species following this catastrophic event almost certainly contributed to its detection.It is even more remarkable that two new species of butterflies (both lycaenids) have now been discovered and described from the ACT and adjacent areas of NSW within the last year, the other being Cyprotides maculosus Braby, 2023, which was first recorded (photographed) only in 2017 (Braby 2023).
Recognition of the newly discovered population of Paralucia in the ACT as a distinct species brings to four the total number of species recognised in the genus.Paralucia crosbyi and P. spinifera are remarkably similar in appearance; the larval and pupal stages are almost identical; and the general biology and life cycle are similar.However, evidence from the adult phenotype and morphology of the immature stages, in particular the egg, indicates that P. crosbyi differs from P. spinifera by a total of 12 morphological character state differences (9 adult stages and 3 immature stages).Perhaps the most striking difference concerns the females of the two species with respect to their dorsal colour pattern, with P. crosbyi being monomorphic and P. spinifera polymorphic, suggesting an underlying genetic difference.Indeed, the two species are phylogenetically reciprocally monophyletic according to molecular data (COI), although the mean number of mitochondrial nucleotide changes separating the two species is small (<1%).In addition, limited samples indicate that the life cycle differs with respect to development, with P. crosbyi larvae (x̄= 78.0 ± 3.65 days) developing substantially slower than those of P. spinifera (x̄= 64.0 ± 6.48 days), at least when reared at room temperature (Table 3).Baker et al. (1993) recorded the larval duration of P. spinifera as varying from 60 to 70 days under laboratory conditions, a range that falls within the range recorded in this study for that species (Table 3).The broad flight season for both species lasts approximately 10 weeks, but adults of P. crosbyi possibly fly earlier in the season, from late July to early October, whereas adults of P. spinifera are recorded to fly mainly from late September to mid-November (Baker et al. 1993).However, more recently, Braby (2000) and Mjadwesch and Nally (2008) have noted that adults of P. spinifera may emerge as early as early September or even mid-August.It is possible that the flight season of P. spinifera has shifted earlier since its initial discovery in the 1970s as a consequence of climate change; however, there is no historical baseline data for P. crosbyi to indicate if the adult phenology of this species has shifted.In 2023, New South Wales (including the ACT) experienced the warmest winter on record, with the mean maximum temperature being 1.72 C above average (Bureau of Meteorology 2023).Thus, it is possible the unusually warm conditions in the ACT may have triggered an earlier adult emergence that season.
Paralucia crosbyi and P. spinifera, together with P. aurifera, form a compact, monophyletic group that is similar in adult morphology, immature stages and ecology.All three species are associated with the same larval food plant, Bursaria spinosa, and the same species of attendant ant, or very similar ants within the Anonychomyrma itinerans (Lowne, 1865) complex or species group, which currently comprises A. itinerans and three subspecies: A. itinerans ballaratensis (Forel, 1902), A. itinerans depilis (Forel, 1902) and A. itinerans perthensis (Forel, 1902).The current state of taxonomy of Anonychomyrma is poor, and taxonomic revision of the genus is required before the species can be identified definitively.The literature lists a different species of ant attending P. aurifera (i.e., Anonychomyrma sp.[nitidiceps species group]) (Cushman et al. 1994;Eastwood & Fraser 1999), but examination of specimens of ants attending P. aurifera held in the ANIC suggests that they are the same as those attending P. spinifera and P. crosbyi (J.Lewis, pers. comm. 2023).
In contrast, P. pyrodiscus is more distantly related and sister to P. aurifera + (P.spinifera + P. crosbyi): the immature stages differ markedly in morphology and are attended by ants in the genus Notoncus (Formicinae) (Braby 1990).Further sampling and taxonomic work is needed to determine if the P. pyrodiscus complex comprises more than a single species given the large genetic divergence observed in this study between the 'bright' and 'dark' phenotypes in New South Wales (4.7% for mtDNA) and those recorded by Roitman et al. (2017) between the subspecies P. pyrodiscus lucida Crosby, 1951 ('bright' phenotype) and nominate P. pyrodiscus pyrodiscus ('dark' phenotype) in Victoria (4.5% for mtDNA).Interestingly, P. crosbyi and P. spinifera occur in similar environments-both are restricted to montane eucalypt open woodland or woodland (typically >900 m) in the lower rainfall zone on the western (inland) slopes of the Great Dividing Range, and adults begin to emerge in late winter/early spring when temperatures are still cold.Populations of P. spinifera typically occur at the edges of woodland or within areas of the habitat subject to regular or episodic disturbance, usually on west-to north-west-facing slopes at altitudes between 850 and 1250 m (Baker et al. 1993;Edwards & Common 1978;Mjadwesch & Nally 2008).Both species have exceedingly limited geographical distributions, yet they are separated by only $250 km.Lack of differences in the male genitalia between P. crosbyi and P. spinifera indicate that either the two species have speciated very recently or there has been a lack of selection reinforcing interspecific mating barriers between them, probably due to allopatry.The low level of mean genetic divergence between these species for COI (0.93%) also suggests that the lineages have diversified very recently.Based on a sequence divergence rate of 2.3% per million years for COI for butterflies (Brower 1994), mean genetic differentiation of 0.93% suggests that the common ancestor of the two species diverged ca.400 000 years ago (Pleistocene).The two species may be vicariant relicts of an ancestral species that was formerly more widespread during cooler episodes (e.g., glacial cycles of the Pleistocene), but the populations have since contracted to higher altitudes that are now disjunct from each other, leading to allopatric speciation.
Further investigations in similar environments in north-eastern Victoria and north-eastern New South Wales (e.g., Ebor Plateau) should be made to determine if there are additional relict populations of this species group that have been overlooked.Searches undertaken on the drier western slopes of Barrington Tops, NSW (between 900 and 1200 m asl) and Coolah Tops, NSW (1080 m), however, failed to detect the presence of Paralucia species other than P. aurifera-five sites comprising montane grassy open woodland or woodland that supported an abundance of Bursaria spinosa were sampled over 7 h during 4-7 September 2023.
Future survey work is needed to investigate the spatial distribution, critical habitat and conservation status of P. crosbyi, given that it appears to be a narrow-range endemic.The butterfly was discovered $19 months after the Black Summer 2019-2020 megafire storms (Legge et al. 2022;Wintle et al. 2020), which in Namadgi NP, started on 27 January 2020 and burned the majority of the park ($80% or 82 700 ha) (Environment Planning and Sustainable Development Directorate 2020).Nearly all known sites of the species were severely impacted by the fire (S.Bond, pers. comm. 2023).During the following season in the spring of 2021, habitats of burnt sites were in an early stage of ecological succession, and the larval food plants were small in stature (<1 m high) and regenerating from basal rootstock; however, adults were exceptionally abundant, indicating that the species had survived the fire and possibly benefitted from this disturbance event.Presumably, the species consists of a single large metapopulation that comprises a series of subunits that are patchily distributed in the landscape of Namadgi NP and the adjacent ranges east of the ACT border.However, it is not yet clear which sites comprise core (permanent) breeding areas or ephemeral (temporary) satellite breeding areas that represent short-term responses to post-fire conditions following the January 2020 fire storm.Monitoring is also required to determine how the population size changes over time as the vegetation returns to a climax state under pre-fire conditions following ecological succession.New et al. (2000) and Nally (2003) observed that Bursaria spinosa is an ecological resprouter following disturbances, such as fire.Nally (2003) noted that following an intense fire in December 1997, the population of P. spinifera at Lithgow increased dramatically but then declined 3 years post-fire.

ACKNOWLEDGEMENTS
This paper is dedicated to the late Ted Edwards (1945Edwards ( -2023) ) for his encouragement, guidance and mentoring with the taxonomy of Australian butterflies over many decades.Ted published a pioneering study on the taxonomy of Paralucia in which he and Ian Common described the sibling species P. spinifera.I am especially grateful to Andreas Zwick for providing the mitochondrial sequences and to Ethan Beaver for assistance with the phylogenetic analysis; Rod Eastwood also provided considerable advice with molecular data.Paul Sunnucks generously made available his mitochondrial sequences of P. pyrodiscus.I am grateful to Jon Lewis for identifying the attendant ants of Paralucia aurifera, P. spinifera and P. crosbyi; David Albrecht for assistance with plant identifications; Suzi Bond and her citizen scientist colleagues for sharing data on P. crosbyi in the ACT; Susan Wishart for making the initial discovery of the new species; Peter Samson for permission to include his photos of the first instar larvae of Paralucia spp.; and Ethan Beaver for assistance with field work.Rod Eastwood, Chris Müller, Simon Nally, Suzi Bond and Jon Lewis kindly reviewed and improved a draft of the manuscript.Material was collected under permit PL2017113 under the ACT Government Nature Conservation Act 2014.Open access publishing facilitated by Australian National University, as part of the Wiley -Australian National University agreement via the Council of Australian University Librarians.
; head black, hidden beneath prothorax; body cream, with short colourless lateral setae; prothoracic and anal plates shining black; prothorax with six long colourless anterior setae; prothoracic plate with two long colourless dorsal setae; mesothorax, metathorax and abdominal segments 1-6 each with two long black dorsal setae, recurved posteriorly; abdominal segments 9-10 with six long colourless setae.
; head dark brown; prothorax grey-green, with prominent black middorsal mark anteriorly and cream dorsolateral mark posteriorly; prothoracic plate grey, mottled with black dots and black lateral markings; mesothorax, metathorax and abdominal segments 1-7 grey-green, with broad cream dorsal band enclosing prominent dark grey-brown middorsal line and reddish subdorsal line angled obliquely or laterally towards F I G U R E 6 First instar larva of Paralucia spp.: (a-d) P. crosbyi sp.nov., showing dorsal and lateral views after hatching (a, b) and dorsal and dorsolateral views before moulting (c, d); (e) P. spinifera lateral view; (f) P. aurifera dorsolateral view.Photo credits: (a, c, d) © M.F.Braby, (b, e, f) © Peter R. Samson.
January 2020 megafire storm, and much of the overstorey had been obliterated by the fire.When the butterfly was first discovered at this site ($19 months post-fire), the larval food plants were approximately 0.5-1.0m high and regenerating from basal rootstock.However, three seasons post-fire (summer of 2022/23), the food plants flowered and set seed.The larval food plant keys to B. spinosa ssp.lasiophylla, but as Cayzer et al. (1999) pointed out, F I G U R E 9 Habitat of Paralucia crosbyi sp.nov. at Namadgi NP, ACT: (a, b) unburnt montane eucalypt open woodland at Nursery Creek Tk, Orroral Valley; (c, d) montane eucalypt woodland at Brandy Flat FT in early successional stage following the January 2020 megafire storm, with larval food plants in foreground; 20 months post-fire (c), 34 months post-fire (d); (e) larval food plant Bursaria spinosa ssp.lasiophylla, note extensive larval grazing maintaining prostrate form; (f) attendant ant Anonychomyrma sp.(itinerans species group).

F
I G U R E 1 1 Life cycle of Paralucia crosbyi sp.nov.showing the phenology of the adult and immature stages.T A B L E 3 Duration of the larval stage (from egg hatching to pupation) of Paralucia spp.reared at room temperature (22 ± 2 C).Samples of eggs of both species were collected in the field in late September 2022 and reared on Bursaria spinosa under identical conditions.
GenBank accession numbers for the mitochondrial COI samples of Luciini analysed in this study.
T A B L E 1 Maximum likelihood phylogeny of Paralucia species inferred from the mitochondrial COI gene or 'barcode' fragment.Numbers adjacent to nodes are SH-aLRT values (left) and ultrafast bootstraps (UFBoot) values (right).