A New Subspecies of Philiris diana Waterhouse & Lyell, 1914 (Lepidoptera: Lycaenidae) from the Wet Tropics of Northern Australia

. Philiris diana Waterhouse & Lyell, 1914 from the Wet Tropics of northeastern Queensland was previously thought to be monotypic, being most closely related to Philiris papuanus Wind & Clench, 1947 from Cape York Peninsula, Australia, and mainland New Guinea. However, a new subspecies was recently discovered on the Atherton Tableland, which we illustrate, diagnose and describe as Philiris diana fortuna ssp. nov. It appears to be a narrow-range endemic, restricted to montane forest (750–1,090 m asl) and allopatric or parapatric from the nominotypical subspecies Philiris diana


Introduction
Philiris Röber, 1891 is a large genus of lycaenid butterflies occurring in Australia and mainland New Guinea and its adjacent islands. It is closely related to Hypochrysops C. & R. Felder, 1860 in the Hypochrysops section of the tribe Luciini (Eliot, 1973), but unlike most related genera in that tribe, the ventral patterns of Philiris species are relatively unmarked and they typically exhibit a silvery-white underside ground colour. Parsons (1998) and Braby (2000) noted that the genus, at that time, contained approximately 65 species.
Since then, eight more species have been described or recognized, mainly from New Guinea and nearby islands (Müller, 2014;Sands, 2015;Tennent, 2016). Only nine species are currently recognized from Australia (Sands, 2015;Braby, 2016), most of which are restricted to eastern and northeastern Queensland. The majority of species are associated with tropical rainforest or rainforest regrowth (pioneer vegetation). Parsons (1998), following the classifications of Tite (1963) and Sands (1979Sands ( , 1981a, proposed several informal species groups within the genus according to various morpho logical characters. One such group, the diana species-group, was distinguished in being relatively large in size, without a black spot on the inner margin of the hindwing underside in both sexes, and usually long, asymmetric valvae in the male genitalia. The diana species-group consists of seven species, viz: Philiris gloriosa (Bethune-Baker, 1908); P. diana Waterhouse & Lyell, 1914 (Figs 4-6, 8-10); P. violetta (Röber, 1926); P. papuanus Wind & Clench, 1947;P. praeclara Tite, 1963;P. montigena Tite, 1963;and P. siassi Sands, 1981. Monophyly of the diana speciesgroup and phylogenetic relationships of its constituent species have yet to be established, although a preliminary molecular phylogeny indicates that five of these species comprise a monophyletic group, while two species (P. montigena and P. violetta) are more distantly related (N. E. Pierce et al., unpublished data). Only P. diana and P. papuanus kerri Sands, 2015 occur in Australia, which have allopatric distributions in northern Queensland. Philiris papuanus kerri occurs widely on Cape York Peninsula, where it extends from Bamaga (Monteith & Hancock, 1977) south to the Chester River area and the Rocky River (Valentine & Johnson, 1997). In contrast, P. diana is restricted to the Wet Tropics, where it is has mainly been recorded from a small number of locations in the Cairns-Kuranda district (Braby, 2000), with the majority of specimens captured from Kuranda early last century by F. P. Dodd. Philiris diana has also been recorded further south in montane areas at Paluma (Common & Waterhouse, 1981) and the Bluewater State Forest (Dunn & Dunn, 1991). However, apart from an anomalous specimen from Lake Eacham (Braby, 2000;Sands, 2015), which we discuss later, we know of no intervening records between the Cairns-Kuranda district and the Paluma Range.
In March 2021, a male Philiris was captured by one of us (BSH) in montane rainforest at c. 900 m above sea level (asl) near Millaa Millaa on the Atherton Tableland. The specimen resembled P. diana in size but there were significant wing pattern differences. Further specimens comprising both sexes were captured during the following wet season. In addition, eggs were obtained from a captive female in early December 2021 and reared to adulthood, and other adults were reared from larvae collected from foliage of the larval foodplant Litsea leefeana (F. Muell.) Merrill (Lauraceae) (Petrie & Hacobian, 2022). Close examination and comparison of this material, and of two females from other locations on the Atherton Tableland, have revealed that they are indeed P. diana but are sufficiently differentiated to warrant subspecific status. Hence, the purpose of this paper is to describe this new population (P. diana ssp. "Atherton Tableland") and to compare it with both nominotypical P. diana and P. papuanus.

Materials and methods
Character differences between the phenotypes of P. diana and P. diana ssp. "Atherton Tableland" were assessed by examination of adult specimens in both museum and private collections. Quantitative measurements were made for the following five morphological characters concerning wing shape and colour pattern elements (Figs 1, 2). To account for adult size variation and possible allometric scaling, all measurements were divided by wing area or wing length (as a proxy of body size) to yield dimensionless quantities. Photographic images of individual specimens (with differing scale factors) were used as the basis for all measurements utilizing SketchAndCalc software (Dobbs, 2011). Sample sizes for each sex were as follows: 46 males (24 P. diana, 22 P. diana ssp. "Atherton Tableland"), and 42 females (26 P. diana, 16 P. diana ssp. "Atherton Tableland"). Data for each character were analysed using a parametric two-sample t-test.
1 The relative size of the white central patch on the upperside of the forewing of females was calculated for each specimen by measuring the area of the white central patch (character B) and then dividing that measurement by the total forewing area (character A). 2 The degree of curvature of the forewing termen of females was assessed by calculating the "curved area", that is, the area of that portion of the forewing enclosed between the termen and a straight line between the end of veins R 5 and 1A+2A (character C). This area was then divided by character A. 3 The degree of curvature of the forewing termen of males was assessed by calculating the quantity character C/A for each specimen, as described above for the female. 4 The relative width of the black terminal band on the upperside of the forewing of males was calculated by measuring the width of the band at vein M 3 (character D) for each specimen divided by the length of the forewing, measured as the straight-line distance between the wing base and the end of vein M 3 (character E). 5 The relative width of the black terminal band on the upperside of the hindwing of males was calculated similarly to the forewing outlined above, with measurements made on vein M 3 (character F) for each specimen divided by the length of the hindwing, measured as the straight-line distance between the wing base and the end of vein M 3 (character G).
Character differences between the early stages of the two taxa were also assessed by examining and comparing the eggs, larvae and pupae of P. diana ssp. "Atherton Tableland" with those of P. diana collected from near Cairns and reared by us (Petrie & Hacobian, 2022). In order to assess the possible influence of ambient temperature on the adult phenotypes of P. diana, larvae from eggs laid by a captive female of P. diana ssp. "Atherton Tableland" in December 2021 and other larvae collected from the field during January 2022 were divided into two cohorts. One cohort was reared at an altitude of c. 920 m asl under near-ambient conditions of temperature (and humidity) approximating those applicable to wild populations ( Table 1). The other cohort was raised near sea-level under a combination of controlled temperature (and humidity) conditions more like those of wild populations of nominotypical P. diana. In addition, two cohorts of P. diana collected from the field near Cairns in December 2021-February 2022 were reared under the same reciprocal environmental conditions. In both experiments, the larvae were reared on the larval foodplant Litsea leefeana (Petrie & Hacobian, 2022).
Dissection of the genitalia was performed using standard techniques: the entire abdomen was removed and placed in 10% KOH and boiled for 20 minutes, 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 then placed in an alcohol-based gel in an excavated glass block and photographed using a Leica M205A microscope and were stacked using Helicon Focus 5.3 according to the technique of Su. Terminology for genitalia follows Klots (1970) and Eliot (1973).
To test if the midland and upland populations of P. diana represented two separate species according to neutral genetic markers, patterns of monophyly and the extent of genetic divergence were analysed for six samples for the "barcode" region of the mitochondrial gene cytochrome c oxidase subunit I (COI), a widely used marker for species level discrimination (Hebert et al., 2003a,b). Three samples of P. diana diana collected from near Cairns in 2022 and a single sample of Philiris diana ssp. "Atherton Tableland" collected from near Millaa Millaa in March 2021  were sequenced (GenBank accession numbers: Philiris diana diana-SUB12522691 MFB-20-C114 OQ214188, SUB12522691 MFB-20-C113 OQ214189, SUB12522691 MFB-20-C112 OQ214190; Philiris diana fortuna-SUB12522691 31-036203 OQ214191). In addition, partial sequences (550-658 bp) were downloaded for two samples on BOLD (Ratnasingham & Hebert, 2007): a female P. diana diana collected from Paluma in 1969 (accession number: ANICS1826-11), and a female Philiris diana ssp. "Atherton Tableland" collected from Lake Eacham in 1998 (accession number: ANICS1824-11). For fresh material, DNA was extracted from one or two legs and then the entire mitochondrial genome was sequenced according to the

Results
Results for the analysis of five quantitative morphological wing characters are presented in Table 2. Three characters were found to be significant among the two different P. diana populations, whereas two traits (male forewing termen curvature and male forewing black terminal band at vein M 3 ) were found to be not significant.  (Table 1), revealed no environmental influence on adult phenotype (Tables 3, 4). Although our samples were small and not amenable to statistical analysis, all adults obtained from immature stages of the montane population P. diana ssp. "Atherton Tableland" (n = 25) reared in either Palm Cove, Cairns (n = 4) or Beatrice, Atherton Tableland (n = 21) produced the montane form P. diana ssp. "Atherton Tableland" (Table 3). Conversely, all adults obtained from immature stages of the lowland population of P. diana (n = 7) reared in either Palm Cove, Cairns (n = 4) or Beatrice, Atherton Tableland (n = 3) produced the typical P. diana form (Table 4).
Mt DNA barcodes (COI) revealed negligeable divergence between the two populations, and the pairwise distances estimated using the K2P model were the same as those for uncorrected p-distances. The average pairwise distance within the four P. diana diana samples was 0.35%, but for the two samples of P. diana ssp. "Atherton Tableland" it was 0%. Although P. diana comprised a monophyletic lineage, the two samples of P. diana ssp. "Atherton Tableland" were nested within the four samples of P. diana diana, indicating lack of reciprocal monophyly between the two taxa ( Fig.  3). The average p-distance between P. diana diana and P. diana ssp. "Atherton Tableland" was 0.28%, with the range varying from 0.15-0.46%. Although the mean level of divergence within P. diana diana was greater than between P. diana diana and P. diana ssp. "Atherton Tableland", these differences were not statistically significant because they were less than 1 bp on average for the COI barcode region (P. diana diana: 0.35% * 658 bp = 2.3 bp; P. diana ssp. "Atherton Tableland": 0.28% * 658 bp = 1.8 bp).        diana males, the upperside of the forewing has a distinct white central patch, which extends below vein CuA 2 , whereas in P. diana fortuna this white patch is generally absent or occasionally represented by a few scattered grey or greyishwhite scales only; this patch, when present, does not extend below vein CuA 2 . (5) The shape of the forewing termen is more strongly arched in P. diana fortuna females, and this difference is statistically significant ( Table 2). (6) The white central patch on the upperside of the forewing in P. diana fortuna females (range: 20-35% of wing area) is smaller compared to that of P. diana diana females (range: 30-40% of wing area), and this difference is statistically significant ( Table 2). (7) On the upperside of the hindwing, the costa and apical areas between veins Rs and M 1 are extensively suffused white and the adjoining central area distal to the discocellular veins between M 1 and M 3 is also white in P. diana diana females, whereas in P. diana fortuna females these areas are substantially reduced in extent and grey in colour or grey with a few white scales. A consequence of the reduction of the white areas in both fore-and hindwings is that the female of P. diana fortuna is a much darker butterfly with more extensive areas of brown-black than the female of P. diana diana. (8) The underside ground colour, in reared specimens of both sexes, is silvery-white in P. diana fortuna, but white with less silver tone in P. diana diana.
The male genitalia of P. diana fortuna  are similar to those of P. diana diana (Figs 21-23), particularly with respect to the shape of the phallus, sociuncus, brachia, and valvae which are asymmetrical in profile. The only difference lies in the width of the long, narrowed middle section of the valvae-in P. diana fortuna the middle sections of the right and left valvae are slightly broader than in P. diana diana. The genitalia of the two P. diana subspecies are similar to those of the two subspecies of P. papuanus-P. papuanus kerri and P. papunus papuanus (Figs 24-29)-in that they show the same degree of asymmetry among the valvae, but in P. papuanus the curved apical spine of the left valva is substantially longer and more robust, and the middle section of the left valva is more strongly arched (in lateral view) than in P. diana. The apex of the right valva terminates in a short beak-like projection that is oriented dorsolaterally in both subspecies of P. diana, whereas in the two P. papuanus subspecies the apex bears a longer but narrower dorsolateral projection that terminates in an outwardly curved spine.

Description
Male. Head: eyes brown when alive, black when dead, ringed by white scales; labial palpus dark grey dorsally, white ventrally, clothed in piliform scales, second (middle) segment four times longer than third segment; antennae 8.7-9.3 mm long, flagellum with 32-35 segments (shaft 18-20, club 14-15), shaft black prominently ringed with white, club black variably marked ventrally with orange-brown most extensively in apical half; frons dark brownish-grey; chaetosemata predominantly grey. Thorax: dorsal surface dark grey, ventral surface white; legs white, marked with black between segments, tarsi and mid tibiae. Forewing: length 15.2-17.8 mm (x̄ = 16.7 ± 0.66 mm, n = 19), upperside dark purplish-blue, costa narrowly edged with black, a narrow black terminal band tapered along its length and broadest near apex, end of veins black, an obscure central area between veins M 3 to CuA 2 suffused dull grey, terminal scale fringe black near apex but elsewhere black with white tips; underside uniform silvery-white, costa narrowly edged with black, terminal scale fringe black near apex but elsewhere white; costa and termen slightly convex, apex pointed but with outer costa distinctly rounded, tornus rounded, dorsum straight. Hindwing: upperside dark purplish-blue, costal region mid grey, apex dark grey with a few scattered white scales, a narrow black terminal band of constant width between veins M 1 and CuA 2 but broader towards apex and tornus, dorsum broadly grey, terminal scale fringe black near ends of veins M 3 , CuA 1 , CuA 2 and 1A+2A, elsewhere black with white tips; underside uniform silvery white, black terminal spots at ends of veins M 3 , CuA 1 , CuA 2 and 1A+2A, a narrow black terminal line joining spots between veins CuA 1 and 1A+2A, terminal scale fringe black adjacent to terminal spots, elsewhere white. Abdomen: dorsal surface dark grey, ventral surface white.
Female. Head: eyes, labial palpus, antenna and frons similar to male; antennae 8.7-9.1 mm long, flagellum with 34-38 segments (shaft 18-21, club 14-17). Thorax: similar to male. Forewing: length 16.4-18.7 mm (x̄ = 17.6 ± 0.73 mm, n = 13), upperside dark brownish-black, with a prominent white central patch edged with suffusion of iridescent blue scales, blue suffusion more extensive towards base of white patch, sometimes a few white scales below central white patch between vein 1A+2A and dorsum; terminal scale fringe black near apex, elsewhere black with white tips; underside similar to male. Hindwing: upperside dark brownish-black, costal and subapical region above vein Rs grey with scattered white scales, sometimes extending to subterminal region below vein Rs, sometimes a few scattered blue scales present in discal cell and subterminal region between veins M 1 and M 3 , dorsum grey, terminal scale fringe black near ends of veins M 3 , CuA 1 , CuA 2 and 1A+2A, elsewhere black with white tips; underside similar to male. Abdomen: similar to male.
Variation. Males vary in several traits, including the hue or tone of the dorsal blue colouration, which ranges from dark cobalt-blue to dark purplish-blue; and the extent of the dorsal greyish central suffusion on the forewing, which is usually absent or rarely present as a very obscure patch; in one individual (5% of specimens examined, n = 22) (Fig. 14) it was far more extensive and overlaid with a few greyishwhite scales, similar to that of P. diana diana. In females, the size of the dorsal white central patch on the forewing shows minor variation, and the white scales may occasionally extend distally into the subterminal area or below vein 1A+2A towards the dorsum; overall, the extent of the white patch varies from 20-35% as a proportion of the wing area. In addition, the extent of the dorsal white scales of the hindwing costa and apex is variable in females, and it may obscure the apex entirely or extend below vein M 1 , as does the degree of blue scales between M 1 and M 3 , which may be absent.

Remarks
Philiris diana was originally described by Waterhouse & Lyell (1914) based on 17 specimens (10♂, 7♀) from Kuranda, QLD, all collected by F. P. Dodd; however, they did not refer to a type of any sort. Peters (1971) referred to a "holotype" male (registration numbers AMS KL.21453 and K.191300) and an "allotype" female (AMS KL.21455 and K.584428) in the Australian Museum, Sydney. Edwards et al. (2001) regarded Peters' incorrect reference to a holotype as a valid lectotype designation. It therefore follows that the subsequent lectotype designations by both Parsons (1998) and Sands (2015) are invalid and thus do not constitute formal nomenclatural acts.
Of the type series in AMS, we have examined and identified the lectotype male (AMS K.191300) and 13 paralectotypes (7♂, 6♀). Waterhouse and Lyell (1914) illustrated two of their syntypes: the lectotype male on pl. 15, figs 270, 271, and a paralectotype female on pl. 13, fig.  183, each in black and white. The lectotype male (Figs 4, 5, 7) and paralectotype female (Figs 8, 9, 11) referred to by Peters (1971) and originally illustrated by Waterhouse and Lyell (1914) are both illustrated here for comparison with P. diana fortuna. The lectotype male is slightly unusual compared with the rest of the type series and other material we have examined in ANIC, NMV and SAMA in that it has a much broader black apical band on the forewing. Newly emerged adults of a male (Fig. 6) and a female (Fig. 10) of P. diana diana from the Cairns district reared by us are illustrated for comparison; they agree with Waterhouse and Lyell's concept of diana considering that the types are 115 years old. Braby (2000) noted that P. diana had been recorded from the Atherton Tableland at Lake Eacham (760 m) based on an anomalous female in the ANIC that was collected at night from a light sheet (E. D. Edwards, pers. comm. 1998). Subsequently, Sands (2015, figs 7, 8) illustrated this specimen and noted that "Females of P. diana are very variable in the extent of white areas on the upperside of both wings…" (Sands, 2015, p. 222). However, close examination of this specimen confirms that it is P. diana fortuna. Further examination of museum material revealed another female, from Lake Tinaroo (750 m) in SAMA. Thus, the concept of the P. diana diana female needs to be reconsidered-females only show slight variation in the extent of the white patches on both wings.

Etymology
The subspecific epithet is derived from the Latin word fortuna, which means chance or luck. This name reflects the part played by luck in the discovery of this subspecies. The first specimen collected by us was a crippled male that could not fly which was found walking on the ground of a vehicle track through rainforest; the second specimen was a female captured whilst laying eggs on foliage of a rainforest tree, thereby enabling documentation of the early stages and confirmation of the larval foodplant.

Distribution and habitat
Philiris diana fortuna is currently known only from the Atherton Tableland in the Wet Tropics of northeastern Queensland. It has been recorded from Lake Tinaroo, Lake Eacham, Beatrice near Millaa Millaa and Millaa Millaa lookout at altitudes between 750-1,090 m asl. All sites occur in upland and montane tropical rainforest where the larval foodplant Litsea leefeana occurs as a mediumsize tree reaching the rainforest canopy or along edges of smaller patches of rainforest regrowth. Within the known extent of occurrence, the foodplant is a relatively common component of regenerating rainforest following disturbance. All known occurrence records of Philiris diana fortuna are from vegetation communities growing on basalt soils.

Biology
The larval foodplant and morphology and colour pattern of the immature stages of P. diana fortuna (Figs 30-35) are fundamentally the same as those of P. diana diana. The only observable differences concern the ground colour of instar III, which is always green in P. diana fortuna but sometimes yellow in P. diana diana, and in the colour of the reddish dorsal blotches in the early instars, which are deep red-brown in P. diana fortuna but often pinkish in P. diana diana. The larvae of P. diana fortuna undergo six instars, like P. diana diana (Petrie & Hacobian, 2022) and Philiris ziska titeus D'Abrera, 1971 (Samson & Johnson, 2009). The pupae are usually more heavily striped in P. diana diana, although the degree of pale brown lateral bands or stripes on the wing case and abdomen is variable in P. diana fortuna and our samples of the former taxon are limited. In P. diana fortuna, the extent of red-brown blotching on segments T1-A2 is variable in larval instars II-IV, and the colour of the dorsal longitudinal band varies from yellow-brown to reddishbrown in instars V and VI.
Males have been observed perching on sunlit leaves during the early and late afternoon, between 1230-1400 h and again c. 1700 h AEST, in late March in the canopy of trees of the larval foodplant Litsea leefeana and other adjacent tree species at heights of 10-25 m above the ground. An adult female was observed, and subsequently captured, perched and basking on a leaf in the lower canopy of the foodplant growing in a small patch of regenerating rainforest approximately 5 m above ground level during the early afternoon (c. 1300 h). Another female was observed at a similar time (1320 h) being pursued by two or three males around sunlit foliage of Flindersia brayleyana F.Muell. growing adjacent to the foodplant. The female was subsequently found in copulation with one of the males (Figs 14, 37) settled on foliage c. 10 m above ground level close to the original pursuit. The pair was captured and remained in copulation for 35 mins until they finally separated at c. 1400 h. Females (Fig. 36) have been observed ovipositing on dried skeletonized patches on the abaxial surface of mature leaves of the foodplant, as low as 1 m above ground level. Hatched eggs and larvae have been found in similar situations. Eggs were never found on soft new growth or juvenile leaves. Larvae were never observed to be attended by ants. However, pupae were noted to stridulate when exposed to strong light. A pupal exuvium was found on the abaxial surface of a partly eaten leaf, but no live pupae were located in the field. In captivity, all but one larva reared (n = 28) pupated on the abaxial surface of the leaf of the foodplant; the single exception pupated on the adaxial surface just above the junction with the petiole.
In terms of phenology and duration of the immature stages, the following developmental times were recorded in captivity during the wet season in December-March: egg 7-10 days; larva 51-82 days (instar I 6-7 days, instar II 6-8 days, instar III 7-9 days, instar IV 8-12 days, instar V 9-15 days, instar VI 15-31 days); pupa 10-14 days. Total developmental time from egg to adult varied from 2-3 months. Adults were collected from the field in November, December, March, and April. Some adults reared from eggs (laid by a captive female in early December) emerged in late February, and three adults reared under laboratory conditions from larvae collected in April emerged in mid-July. One larva, collected from the field in early June, was reared under unheated conditions and pupated in late October. Thus, the broad flight period, as currently known, extends from November to April, with at least two generations completed annually.

Discussion
In their original description of P. diana papuanus, Wind & Clench (1947, p. 6) admitted that "there is more than a possibility that papuanus may be a species in its own right." Sixty-eight years later, Sands (2015) treated P. papuanus as a distinct species with two subspecies: P. papuanus papuanus Wind & Clench, 1947 from mainland New Guinea, and P. papuanus kerri Sands, 2015 from northern Australia. In our study, fundamental differences in the male genitalia observed between P. diana and P. papuanus, particularly the shape of the asymmetric valvae, confirm the two species hypothesis recently proposed by Sands (2015) for this set of taxa based mainly on adult phenotype. Tite (1963, p. 235) stated that "The male genitalia [papuanus] are basically the same as those of the nominate race [diana]", although his illustrations (figs 83-84, p. 238) clearly show differences in the morphology of apical spine of the left valva. Sands (2015) also drew attention to this character difference between P. diana and P. papuanus. Within P. papuanus, negligible differences were observed in the genitalia between P. papuanus papuanus and P. papuanus kerri, supporting the subspecific classification proposed by Sands (2015). Philiris diana and P. papuanus are also differentiated according to larval foodplant specialization and morphology of the immature stages (Wood, 1984;Parsons, 1998;Petrie & Hacobian, 2022). The larval foodplant of P. diana is Litsea leefeana (Petrie & Hacobian, 2022; this study), whereas for P. papuanus it is L. breviumbellata C. K. Allen in Australia (Valentine & Johnson, 1997) and L. guppyi F.Muell. ex Forman in Papua New Guinea (Parsons, 1998). Wood's (1984) reference to L. leefeana as the foodplant for both P. papuanus and P. fulgens kurandae Waterhouse, 1903 from Iron Range is erroneous and is actually L. breviumbellata (G. A. Wood, pers. comm. 2022). The known distribution of L. leefeana does not extend to Cape York Peninsula, its northern-most limit being Cooktown (AVH, 2022).
Within P. diana, there are considerable differences in the wing colour pattern elements and one morphological character (total of 8 characters) between the low to midaltitude and the upland-montane populations; however, the male genitalia, larval foodplant and life history, including morphology of the immature stages, of P. diana fortuna are fundamentally the same as those of P. diana diana. The habitat preferences of adults of the two taxa also appear to be very similar based on our limited observations. Despite substantial differences in adult phenotype, lack of differentiation in other traits (morphology, ecology), as well as lack of reciprocal monophyly and negligible (< 0.3%) pairwise divergence in mitochondrial COI sequences between the two taxa, suggests subspecific classification is the most appropriate taxonomic rank for the montane population of P. diana according to the criteria of Braby et al. (2012). Limited rearing of the immature stages of each subspecies under different environmental conditions produced identical phenotypes for each respective subspecies (Tables 3, 4). Although further experimentation is needed to confirm this pattern, the results do suggest the distinct phenotype of each subspecies is under genetic control rather than environmental influences.
The Wet Tropics biome of northeastern Queensland, stretching from Cooktown to Townsville, is noted for its high level of species richness and endemism, including invertebrates (Yeates et al. 2002;Yeates & Monteith, 2008). The region is also exceptionally rich for butterflies (Kitching, 1981;Kitching & Dunn, 1999), with around 225 species or 57% of the butterfly fauna recorded from the Australian mainland (Braby, 2000). Braby & Müller (2014) also noted that the region was an important area of endemism for rainforest butterflies, with five species and 14 subspecies found nowhere else. At that time, P. diana was considered to be a polytypic species that included P. papuanus from mainland New Guinea and Cape York Peninsula. Following the taxonomic revision of Sands (2015), P. diana must now be considered a narrow-range endemic and thus one of a set of six butterfly species endemic to the Wet Tropics.
The discovery of P. diana fortuna further highlights the importance of the Wet Tropics as a biodiversity hotspot. This taxon appears to be restricted to the Atherton Tableland where it occurs in upland and montane forest between 750-1,090 m a.s.l. The Atherton Tableland has been identified as a key upland biogeographic unit (the Atherton Uplands subregion) noted for its high level of regional endemism (Nix, 1991;Williams et al., 1996;Yeates et al., 2002). By contrast, P. diana diana is mainly restricted to low to mid-altitudes in the coastal escarpment in the Cairns-Kuranda district between 50-500 m with an outlying record from the Paluma Range at c. 910 m in the Spec Uplands subregion (based on a female in ANIC labelled "4 miles W of Paluma, Q, 3000 ft, 15 Apr. 1969, I. F. B. Common & M. S. Upton"). We have not been able to trace the female specimen recorded from the Bluewater Range (Halifax Uplands) mentioned by Dunn & Dunn (1991). The northern-most occurrence of P. diana fortuna is Lake Tinaroo, whereas the southern-most occurrence of P. diana diana in the Cairns district is near Lake Morris 20-25 km NE of Lake Tinaroo. Thus, the two subspecies of P. diana appear to be allopatric, or possibly parapatric, separated by a minimum distance of only 20-25 km. Presumably, the ancestor of P. diana fortuna became isolated during past glacial cycles of the Pleistocene and its range contracted to montane rainforest refugia in the Atherton subregion (Nix, 1991) after which it differentiated allopatrically from P. diana diana. However, unlike a few other insects (Yeates & Monteith, 2002), its range has not expanded into adjacent upland subregions. Further sampling within the zone of allopatry is required to determine if the two taxa are indeed spatially separated. A study of the phylogeography of P. diana may help determine the extent of historical and contemporary gene flow between the two subspecies and elucidate the timing and process of differentiation.