Resolving a century-old case of generic mistaken identity: polyphyly of Chitoniscus sensu lato resolved with the description of the endemic New Caledonia Trolicaphyllium gen. nov. (Phasmatodea, Phylliidae)

Abstract With every molecular review involving Chitoniscus Stål, 1875 sensu lato samples from Fiji and New Caledonia revealing polyphyly, the morphology from these two distinct clades was extensively reviewed. Morphological results agree with all previously published molecular studies and therefore Trolicaphylliumgen. nov. is erected to accommodate the former Chitoniscus sensu lato species restricted to New Caledonia, leaving the type species Chitoniscuslobiventris (Blanchard, 1853) and all other Fijian species within Chitoniscus sensu stricto. Erection of this new genus for the New Caledonian species warrants the following new combinations: Trolicaphylliumbrachysoma (Sharp, 1898), comb. nov., Trolicaphylliumerosus (Redtenbachher, 1906), comb. nov., and Trolicaphylliumsarrameaense (Größer, 2008a), comb. nov. Morphological details of the female, male, freshly hatched nymph, and egg are illustrated and discussed alongside the Chitoniscus sensu stricto in order to differentiate these two clades which have been mistaken as one for decades.


Introduction
Phasmatodea, the stick and leaf insects, are a group of rather large herbivorous insects that is well known for their diversity of morphological adaptations facilitating imitation of parts of plants (Bedford 1978). One particular lineage of phamids is well known for its perfect mimicry of plant leaves: the walking leaves (Phylliidae). The phylogenetic relationships of phasmids in general Goldberg et al. 2015;Robertson et al. 2018;Büscher et al. 2018a;Glaw et al. 2019;Simon et al. 2019) and phylliids in particular (Bank et al. 2021) are subject to several studies based on morphological, as well as molecular data and corroborated the monophyly of Phylliidae on the one hand but recovered Phyllium as paraphyletic (Bank et al. 2021). Chitoniscus sensu lato in particular are rather short and stout leaf insects, which perfectly blend in with the foliage of the plants they dwell on (Fig. 1). Within their largely tree dominated, canopy dwelling habitats (Fig. 2) such small leaf insects are well hidden from potential predators. Due to their size and their, more or less, close distribution on New Caledonia and the Fiji islands, several species in past years were assigned to Chitoniscus sensu lato based on general morphological features. Large-scale phylogenetic studies on Phasmatodea repeatedly recovered Chitoniscus sensu lato as polyphyletic (e.g., Buckley et al. 2009;Bradler et al. 2015;Robertson et al. 2018;Forni et al. 2020;Bank et al. 2021). Apparently the two distributional centers (New Caledonia and Fiji) of these superficially similar types of leaf insects represent two separate lineages within Phylliidae, with the clade from Fiji always recovered as sister to the remaining phylliids, while the placement of the New Caledonia clade has been recovered within the remaining phylliids in different locations depending upon the tree topology recovered (Buckley et al. 2009;Bradler et al. 2015;Forni et al. 2020;Bank et al. 2021). Convergent traits are common within Phasmatodea (Buckley et al. 2009;Goldberg et al. 2015;Robertson et al. 2018;Büscher et al. 2018Zeng et al. 2020) and particularly insular dwarfism is a common explanation for some small size convergences in vertebrates (e.g., Lomolino 2005) and also reported for insects (Hayashi 1990). Therefore, in light of the phylogenetic results, we aimed to review the assemblage of taxa assigned to Chitoniscus sensu lato to deliver morphological evidence for the two similar sized lineages, that have been considered congeneric since 1904 (Kirby 1904). The distribution, including the furthest latitude south reported for any phylliid, of these miniaturized phylliids indicates two distinct geographic clusters, that render the monophyly of Chitoniscus sensu lato even more unlikely. In turn, the two distributional centers of Chitoniscus sensu lato are not only separated, they are also interconnected by different phylliid taxa (Brock et al. 2021). Hence, the resemblance of the two lineages rather likely arises from secondary adaptation towards similar selective pressures, i.e., convergently. Specimens formerly assigned to Chitoniscus sensu lato were traced in museums, private collections, and other sources and were examined to evaluate the identities of the two groups of small-sized leaf insects. We challenged the present interpretation of Chitoniscus sensu lato with this dataset and compared these two lineages of small sized Phylliidae using morphological and distributional data. In the following we aimed to answer the research questions: i) Which morphological  characters separate these two groups of leaf insects? ii) How can the observed molecular disjunct within Chitoniscus sensu lato be explained phenotypically and geographically? iii) How can this conflict be resolved taxonomically?

Photography
Photographs of specimens deposited within the IMQC collection and Coll SLT were taken by René Limoges using a Nikon D850 DSLR camera (Nikon Corporation, Tokyo, Japan) with Nikon Micro-Nikkor 200mm f/4 lens on Manfrotto 454 micrometric positioning sliding plate (Manfrotto, Casolla, Italy). Lighting was provided by two Nikon SB-25 flash units with a Cameron Digital diffusion photo box (Henry's, Vancouver, Canada).
Photographs from SDEI were taken by Arne Köhler (SDEI) using either a Nikon D7200 camera (Nikon Corporation, Tokyo, Japan) or Leica M205C-microscope (Leica Microsystems Inc., Buffalo Grove, USA). The stacking software used were Zerene Stacker (for the Nikon images; Zerene Systems LLC, Richland, USA) and Helicon Focus (for the Leica images; Helicon Soft Ltd., Kharkiv, Ukraine).
Photographs of specimens within the first authors collection (Coll RC) were taken by RC using a Canon 5D Mark II and a MP-E 65mm macro lens and stacked using Zerene Stacker (Zerene Systems LLC, Richland, USA). The eggs of Trolicaphyllium sarrameaense comb. nov. were photographed using a Nikon D3500 camera (Nikon Corporation, Tokyo, Japan) mounted on a Wild M3C stereomicroscope (Wild, Heerbrugg, Switzerland) and manually stacked. Adobe Photoshop Elements 13 (Adobe Inc., San Jose, USA) was used as post processing software.

Scanning electron microscopy (SEM)
Micrographs were obtained from dried samples, which were sputter coated with 10 nm gold-palladium. Overview images were obtained with a Hitachi TM3000 SEM (Hitachi High-technologies Corp., Tokyo, Japan) at 15 kV acceleration voltage using a rotatable specimen holder (Pohl, 2010). Detailed micrographs were obtained in the SEM Hitachi S4800 (Hitachi High-technologies Corp., Tokyo, Japan) at an acceleration voltage of 5 kV. Adobe Photoshop Elements 13 (Adobe Inc., San Jose, USA) was used as post processing software.

Results
The results of our morphological and biogeographic examination of Chitoniscus sensu lato strongly support the presence of at least two independent lineages, thereby necessitating the recognition of a new genus, Trolicaphyllium gen. nov., to accommodate the endemic New Caledonia taxa. The erection of this novel genus is supported by the identification of autapomorphic morphological features (discussed below) as well as years of molecular results recovering Chitoniscus sensu lato as polyphyletic (e.g., Buckley et al. 2009;Bradler et al. 2015;Robertson et al. 2018;Forni et al. 2020).
Taxonomic hierarchy. Due to the general phylliid morphological features, and the consistent recovery of this clade nested within the greater phylliids in molecular studies, we herein place this genus within the tribe Phylliini Brunner von Wattenwyl, 1893.
Discussion. The selected type species for this new genus is Phyllium brachysoma Sharp, 1898 (= Trolicaphyllium brachysoma (Sharp, 1898), comb. nov.) which was the first species described and is represented by two female syntype specimens collected on Lifou Island (Fig. 22). With the differentiation of the various species within this genus somewhat vague due to possible morphological variability, we felt the original species from a single known exact locality was the best choice as type species.
This new genus has been confused for decades with the similarly sized Chitoniscus Stål, 1875 sensu stricto from nearby Fiji due to their superficial similarities. All molecular phylogenies which have included both Fijian and New Caledonian samples have recovered these as polyphyletic (e.g., Buckley et al. 2009;Bradler et al. 2015;Robertson et al. 2018;Forni et al. 2020;Bank et al. 2021), with the Chitoniscus sensu stricto as sister to all other extant phylliids. Within the phylliid-wide phylogeny of Bank et al. (2021) the New Caledonian clade was recovered as sister to Comptaphyllium Cumming et al. 2019b with high support. Interestingly, few morphological features link these two genera, and it appears as though based upon morphological similarity, higher level relationships among the phylliids are difficult to ascertain. Only the intra-generic relationships appeared to agree readily when reviewing molecular and morphological data (Bank et al. 2021).
Little is presently known about the Trolicaphyllium gen. nov. ecology at the moment, as the only host plant records we have seen to date are from a Ficus sp. (recorded by Thierry Salesne; New Caledonia) and Syzygium cumini (recorded by Sylvie Cazeres (IAC); Fig. 3). The only additional information we have regarding the ecology of this genus are short notes gleaned from specimen labels. In particular, "rainforest" appeared on many labels within the QM collection as noted by Geoff Monteith.
Morphological differentiation from Chitoniscus sensu stricto. Features which liken these two genera together are their short length (ca. 40 to 60 mm) and broad bodies. Both genera have species which have smooth/tapered abdominal shapes or can be strongly lobed (within both males and females). The eggs of both species are small and lack pinnae therefore they superficially resemble each other.
However, when the finer details of these two genera are reviewed, the differences between them are significant (Table 1). A key to genera is not presented here as a thorough key was published within Bank et al. (2021) to all phylliid genera and can still be used to key to the Chitoniscus sensu lato couplet, at which point the features within Table 1 can then be used to differentiate these genera.
Autapomorphic features. Several morphological features unite the New Caledonian species and support monophyly of this clade within the phylliids. Within females, the euplantula 2 and 3 on the tarsus has the unique feature of a ridge-like expansion running along the entire tarsomere (Fig. 12B), a feature not seen in any other phylliids. Within males the alae venation (Fig. 14) is unique within the phylliids as the media anterior (MA) and media posterior (MP) veins fuse with the cubitus (Cu) at different locations along the cubitus and run fused to the wing margin (versus other phylliid Table 1. Summary of morphological features for differentiating Chitoniscus Stål, 1875 sensu stricto from Trolicaphyllium gen. nov. as these two have been mistakenly associated for more than a century. Listed morphologically from the anterior to posterior.
Antennomeres III, VIII, and IX not widened, with similar diameter as the antennomeres between; Fig genera which for example can have the MA and MP often fuse and run together to the wing margin, fuse with the Cu after first fusing together, never fuse and simply fade before reaching the margin, or fuse with the Cu at different locations but are also joined by the first radial (R1) and radial sector (Rs) and all run together to the wing margin).
These autapomorphic features help to define the new genus Trolicaphyllium gen. nov. within the Phylliidae as well as differentiate them from the Chitoniscus sensu stricto. Generic characteristics. The Trolicaphyllium gen. nov. are small to medium, with females ranging from 42.0 mm (in the smallest recorded Trolicaphyllium erosus comb. nov.; Redtenbacher 1906) to 60.0 mm long (in the largest Trolicaphyllium sarrameaense, comb. nov.; Größer 2008b), with males from 38.5 mm to 43.3 mm (in the smallest and largest Trolicaphyllium brachysoma, comb. nov.; Größer 2008b). Typical general coloration is green, but in captivity orange/yellow has been induced (Fig. 19).
Legs. Both sexes have interior tibial lobes on the protibiae which span the full length, lack lobes on the protibial exterior, and the meso-, metatibiae are simple, lacking both interior and exterior lobes. In both sexes the profemoral interior lobe is broader than the exterior lobe (distinctly so in males with a width almost two times that of the exterior lobe, sometimes in females the interior and exterior are almost even in width). In both sexes the profemoral interior lobe is generally only marked with three or four broadly spaced teeth (quite dulled in females; slightly more serrate in males). Both sexes have the interior meso-, and metafemoral lobes slightly broader or about even in width to the exterior lobes, but the interior lobes are always more prominently marked by serration.
Antennae. Females have antennae with nine segments with segments I, III, VIII, and IX notably broader than the other segments ( Fig. 4A) and the stridulatory file has more than 35 teeth (Fig. 5A). Males have antennae which range from 23 to 26 segments with most segments covered in setae which are longer than the segment is wide.
Head capsule. Males have well-developed ocelli (Fig. 13A), and both sexes have head capsules which are marked throughout by distinct granulation which is relatively evenly spaced and, in some cases, appears to be in slightly anterior to posterior rows (Figs 7A, 13A).
Thorax. The thorax is similar in both sexes with mesopleurae that are narrowly diverging from the anterior to the posterior and are marked with five to seven tubercles, occasionally with sparse setae interspersed (Figs 7A, 13A). In both sexes the prescutum is about two times wider on the anterior than long with lateral margins marked by six to eight tubercles, and a prescutum surface which is only slightly granular. When viewed laterally, both sexes have the prescutum anterior rim marked prominently with a raised sagittal spine and both have a prosternum which is prominently marked by a broad, warty tubercle (Figs 8A, 13B).
Wings. Female tegmina are always long, reaching onto abdominal segments VII or VIII and male tegmina are moderate in length, reaching onto abdominal segment III. Females always have highly reduced alae, no more than just a nub (Fig. 22A). Male alae are always fully developed in an oval-fan configuration and reach onto abdominal segment IX (Fig. 14). Female tegmina have a subcoastal vein; radial vein which runs    parallel with the media and splits into the first radial about halfway through its length and terminates in a radial sector and in a small radial to medial crossvein which does fully connect; a bifurcate medial vein; a bifurcate cubitus vein; and a first anal vein which fuses with the cubitus early on (Fig. 10A). Male tegmina have a subcoastal vein; radial vein which runs parallel with the media throughout the full length of the wing and branches into the first and second radial about one third and two thirds of the way through the wing length respectively and terminates as the radial sector; the media runs parallel with the radius and has two media posterior splits near the central area of the wing and terminates as the media anterior; the cubitus is unbranched; and there is a first anal which fuses with the cubitus early on (Fig. 14). Male alae (Fig. 14) have a costal vein running along the anterior margin; a subcostal vein which runs for about two thirds of the length and then fuses with the costal vein; the radial vein is bifurcate when it splits about two fifths of the way through the wing length where they diverge, run parallel, then converge sharply at the apex but don't seem to reach the wing margin; the media is the most unique feature of the alae as it splits early on into the media anterior and posterior which run parallel until the media posterior fuses with the cubitus followed by the media anterior also fusing with the cubitus; the cubitus is fused with the first anterior anal for the majority of the length until the first anterior anal splits and runs to the wing margin; the cubitus, media anterior, and media posterior run fused to the wing margin; the anal veins are split into two groups, the anterior anals and the posterior anals (with seven anterior anals and five posterior anals).
Abdomen. Both sexes have variable abdominal shapes; females can range from spade-shaped to broad and boxy with prominently projecting abdominal lobes VII and VIII; males can be narrowly-ovoid and lack lobes to broadening until segment VII and converging with lobes. Female subgenital plate is short and stout with the apex reaching the anterior margin of the terminal abdominal segment and ending in a fine point; the gonapophyses VIII are long and slender, slightly exceeding the apex of the terminal abdominal segment; the cerci are relatively flat, marked sparsely with a granular surface with margins slightly marked with setae (Fig. 11A). Males have a broad, triangular vomer which is singularly pronged, hooking up into the paraproct.
Egg. Egg morphology is only known at present from Trolicaphyllium sarrameaense, comb. nov. (Fig. 15A-C). Yasumatsu (1942) suggested an egg for Trolicaphyllium brachysoma comb. nov. but the specimen the eggs were from was not collected on New Caledonia and likely represents a different genus and is unrelated to Trolicaphyllium gen. nov. based upon the illustration given. This general egg description is based upon examined material and on images of eggs from several sources all appearing to come  from Trolicaphyllium sarrameaense comb. nov. females. Average length approximately 3 mm long. Eggs when viewed laterally are somewhat rectangular but with the dorsal surface slightly convex and longer than the ventral, giving the egg a slight bent appearance (Fig. 15B). Surfaces are marked throughout with shallow, irregular smooth patches which are accentuated by having darker coloration than the overall egg coloration. Eggs lack pinnae, but instead have small granulation scattered across the capsule which is most prominent and abundant along the capsule margins and notably sparse on the flat surfaces. The egg operculum is conically raised on the ventral margin only, not centrally raised like most phylliid eggs. The raised operculum is only about half as tall as wide and increases from the dorsal margin to the highest point on the ventral margin. The operculum apex has a similar granulation to that found on the capsule margins. Overall egg coloration variable, from a pale tan to light brown, or darker brown, with the pitting on the capsule darker in color and the granulation throughout lighter in color (Fig. 20).
Nymphs. Freshly hatched nymphs are known at present for Trolicaphyllium sarrameaense comb. nov. (Fig. 18A) but are unknown for the other Trolicaphyllium gen. nov. species. Therefore, a comparison between species is not possible at this time. This  generalized description is based upon images of Trolicaphyllium sarrameaense comb. nov. shared by Detlef Größer (Germany). Body long and slender; profemora and protibiae with thin interior lobes but lack exterior lobes; meso-and metafemora with thin interior and exterior lobes; meso-and metatibiae simple, lacking lobes. The base coloration throughout the antennae, head, thorax, abdomen, meso-and metafemora is black. Profemora and all tibiae and tarsi are lighter colored, ranging from dark brown to tan/reddish. All joints between the tibiae and femora are marked with white. The meso-and metafemoral exterior lobes are marked with a medial white spot occupying approximately the central third of the lobe. The abdomen is slender and longer than the antennae, head, and thorax combined. Centrally the abdomen is black, but the margins of segments II-IV and VI-VIII are bordered with a lime green color.

New combinations
Trolicaphyllium brachysoma (Sharp, 1898), comb. nov. Trolicaphyllium erosus (Redtenbachher, 1906), comb. nov. Trolicaphyllium sarrameaense (Größer er, 2008a), comb. nov.  Etymology. Trolicaphyllium meaning "leaf that walks noiselessly". This generic epithet is a compound of the Latinized name Phyllium the type genus for the family (from Greek φυλλον, -ου (phyllon, -oy) + -um; Poitout 2007), coupled with the prefix tro lica from the Drehu (Dehu) language phrase which means "walk noiselessly" (Tyron 1967). We wished to honor the original inhabitants of this area by using a local traditional language. We chose this name because these beautiful insects are so elusive and noiselessly living in the trees where they are often overlooked. This new genus is neuter in gender, following Phyllium.
Distribution. At present Trolicaphyllium gen. nov. specimens are only known from the country of New Caledonia, with records from Grande Terre, Lifou, Tiga, Maré, Ile de Bélep, and L'Île-des-Pins islands (Fig. 21). Likely other islands may also be suitable, but we have yet to locate specimen records from museums or observations. (Sharp, 1898), comb. nov. Figures 14, 22 Fig. 22). See Suppl. material 1 for additional specimens reviewed, their collection data, and depositories.

Trolicaphyllium brachysoma
Remarks. This was the first phylliid species recorded from New Caledonia and was therefore the first described Trolicaphyllium gen. nov. species, consequently, we here designate it as the type species for the new genus. Additionally, it was chosen as it has acceptably accurate collection data (Lifou Island; a rather small island instead of a general locality from the larger main island, which possibly contains several species) thereby removing some degree of possible confusion which could surround such old and difficult to distinguish specimens. This precise locality will allow future reviewers with adequate material sampled from numerous islands to identify species boundaries and determine if this species ranges across New Caledonia or if it is restricted to Lifou Island.
The syntype females were collected by Dr. Arthur Willey in 1897 while he was living on Lifou Island ( Fig. 23; Sharp 1898). Dr. Willey was traveling and living in New   Britain, New Hanover, eastern New Guinea, and Lifou Island between 1895 and 1897 in search of living Pearly Nautilus colonies which he could capture, collect eggs from, and rear through development in order to study their embryology (Wiley 1899; Kerr 1943). Although his years of expedition yielded many great discoveries of which he  . Nymph size is approximated to be relative to each other based upon the few photographs available but is only an estimate A Trolicaphyllium sarrameaense comb. nov.; overall nymph length from head to tip of abdomen approximately 7 mm (Größer 2008b); illustration based upon photographs from Detlef Größer (Germany) B Chitoniscus sp. "Suva'' based upon images supplied by Mayk de Haan (Belgium). published prolifically, he was unfortunately unsuccessful in his primary goal of rearing eggs to maturity (Willey 1899). Willey lived on the west coast of Lifou Island on "Sandal Bay" (modern Santal Bay) from July 1896 to March 1897; and while no exact date was given with the syntype set of females, they are noted as being collected in 1897. Therefore, they are from the beginning of the year (January through March), and most likely from late January when a severe gale passed through the area (wreaking havoc on his Nautili traps; Kerr 1943) which likely knocked the phylliids from the canopy enabling them to be found by Dr. Willey. While males occasionally will fly to lights   at night, females and nymphs are most often only found on the ground after storms when they are knocked from their typical canopy habitat and found lower (Brock andHasenpusch 2003, 2015).
No etymology was given by Sharp, but it can be assumed that he chose brachysoma to denote the size of the species, from the Greek words brachy-(short) and -soma (body).
Differentiation. For female Trolicaphyllium brachysoma comb. nov., one feature which appears to differentiate this species from the other two is the abdominal shape, which is lobeless, and tapered, giving them a spade-shaped appearance. It is worth noting however that in many phylliids abdominal shape is often a poor feature for differentiation as it is often variable within a single species (Cumming et al. 2020b), and even in the syntype set of two females, one female is notably more tapered (Fig. 22B) than the other (Fig. 22A). Trolicaphyllium brachysoma, comb. nov. is about the same size as Trolicaphyllium sarrameaense comb. nov. (ca. 60 mm) which can differentiate them from Trolicaphyllium erosus comb. nov. which are notably smaller (ca. 40 mm).
Correctly matching up male and female phylliids is frequently a significant challenge due to their elusiveness in nature and sexual dimorphism and therefore opposite sexes can only be confirmed through molecular comparison or captive rearing (Cumming et al. 2020c;Cumming et al. 2021). Unfortunately, we have yet to confidently confirm a male Trolicaphyllium brachysoma comb. nov. and therefore, at this time can only illustrate presumed males (Fig. 24) which follow the morphology of the female by having a smooth tapered abdomen and falling within an appropriate size range for a potential male (38 to 43 mm). Trolicaphyllium erosus comb. nov. has no presumed male records we are aware of (as no possibilities have been located in collections which are  small enough to represent a male of this species) but based upon female size the male Trolicaphyllium erosus comb. nov. is likely rather small.
Distribution. The type locality for this species is Lifou island, but brachysomalike specimens with the tapered, lobeless abdomen have been found on Grande Terre (Fig. 21) and L'Île-des-Pins (Fig. 25) as well. Additionally, within the MNHN there is a female which was collected on Ile de Bélep, which is the only phylliid record we have seen from this island, and we only tentatively note this specimen as this species as it has slight lobes on the abdomen and could not be examined in person. Hopefully future molecular analyses with material from multiple islands will reveal if these are all one species or several. (Redtenbacher, 1906  See Suppl. material 1 for additional specimens reviewed, their collection data, and depositories.

Trolicaphyllium erosus
Remarks. This was the second species described from New Caledonia and was described by Redtenbacher (1906) where he gave little to differentiate the species from Trolicaphyllium brachysoma comb. nov. except for the abdominal shape and the overall size as being smaller.
Within the original description by Redtenbacher the number of syntypes was vague, but at least three are explicitly stated as "New Caledonia (Coll. M., Mus. Paris); New Guinea (Mus. Budapest)" (Redtenbacher 1906). At least one syntype was within Redtenbacher's own collection (noted as "Coll. M." within his work) and eventually his collection was deposited in Vienna where it resides today, and the specimen could be traced (Fig. 27; Kaltenbach 2003). Additionally, he states at least one syntype from the "Mus. Paris" which has also survived and was traced (Fig. 26).
An additional syntype was explicitly noted as a nymph within the Budapest Museum (with the wording suggesting that there is only one syntype within that collection). Unfortunately, a fire during the Hungarian Revolution of 1956 destroyed the Budapest syntype along with many of the museum's important type specimens (Brock 1998;   Földvári and Papp 2007;Sabaj 2020). This syntype within the Budapest Museum was noted as being a nymph female from New Guinea with large broad forelegs reminiscent of "Phyllium pulchrifolium Serv." (Redtenbacher 1906). Based on this information it is very likely that this was not actually a Trolicaphyllium gen. nov. specimen as this genus is restricted to New Caledonia and is only known to have rounded profemoral lobes. Instead, this nymph specimen was more likely a female Nanophyllium Redten-bacher, 1906 of which certain species can be small and reminiscent of Trolicaphyllium gen. nov. or Chitoniscus specimens but have broader profemoral exterior lobes (for example Nanophyllium chitoniscoides (Größer, 1992)).
Within the MNHN there is an adult female syntype specimen collected by Philippe François (Fig. 28) from Canal Woodia (canal de Woodin), a common shipping path which runs between Ile Ouen and the south coast of Grande Terre (Bernard 1894). The MNHN type database has the specimen reported as being collected sometime in 1892 (based upon the specimen label "-92"), however, this record number appears to not correspond to the collection year as Francois was not in Oceania 1892. By all accounts, François was actively researching in Oceania from 1888-1891, was back in France in 1892, and then returned to Oceania from 1893-1895(Bulletin Scientifique 1909Dewarumez 2011). We believe that this date on the label instead corresponds to the date the specimens were received by the museum and was instead collected by Francois between 1888 and 1889 when he passed through Canal Woodia (canal de Woodin) several times during his first journey to New Caledonia to study the biology of coral reefs (Bouyssi 1997).
Interestingly, Redtenbacher only appears to have measured one specimen of the type series, the one within his own collection (now within the NHMW), most likely he had seen the Paris and Budapest specimens prior and did not measure them while reviewing them. Harald Bruckner (NHMW) measured the single syntype within their collection and it agrees with the measurements recorded by Redtenbacher (Fig. 27). Interestingly, the smaller overall body length (42.0 mm) appears to be somewhat artificial as the abdomen is significantly excavated and somewhat ballooned outward, pulling the abdomen shorter (Fig. 27C). This is supported by the observation that the other features of this specimen (such as the tegmina length; 27 mm) are actually closer to average sizes for other Trolicaphyllium gen. nov. specimens of other species. Realistically it seems this specimen if it were naturally flat might actually be closer to a range of 50 mm, and not significantly smaller after all.
No etymology was given in the original description, but it can be assumed that Redtenbacher was referencing the lobed abdominal shape with erosus, from the Latin e-(out/away from) and -rosus (gnaw/peck), noting that the leaf-like body appears gnawed on, giving it the undulating appearance. Differentiation.
To date this appears to be the least known species on New Caledonia as we have not been successful in tracing any additional females which fall within this small size range of ca. 40 mm beyond one of the syntypes. All specimens examined so far appear to belong to the other species with their larger size (ca. 50-60 mm). In fact, it is only this smaller size which we have found as useful for differentiation as the other features match up with the prominently lobed Trolicaphyllium sarrameaense comb. nov. females. The additional specimens identified as Trolicaphyllium erosus comb. nov. within the MNHN are larger than the syntype within the NHMW and several are rather morphologically similar to Trolicaphyllium sarrameaense comb. nov. due to their larger size and lobed abdomen but are at this time tentatively left identified as Trolicaphyllium erosus comb. nov. within the collection. Additionally, we have yet to locate any possible male Trolicaphyllium erosus comb. nov. as all males located were much too large and appear to belong to the other two species.
Distribution. The only specimen we have located with detailed locality information is the syntype female from the MNHN collection which has the additional information of "Canal Woodia" which is a small canal between Ile Ouen and the south coast of Grande Terre (Fig. 21). The additional non-type specimens from the MNHN are included within the distribution map, but only tentatively as they are slightly larger than the measured syntype from NHMW and they have variable abdominal shapes, some more strongly lobed than others. (Größer, 2008), comb. nov. Figures 1, 3 See Suppl. material 1 for additional specimens reviewed, their collection data, and depositories.

Trolicaphyllium sarrameaense
Remarks. As it was only described in 2008, this was the most recently described species from New Caledonia with type material originally collected by Sigetake Suzuki in 2004 from Sarramea (Größer 2008a). Within the original description this species was not explicitly compared with the sympatric and morphologically very similar Trolicaphyllium erosus comb. nov. but was instead only differentiated from Chitoniscus lobipes Redtenbacher, 1906, where most features given for differentiation were simply the features we discuss above as significant for differentiating the two genera.
Other lobed specimens have been recovered from throughout New Caledonia, but unfortunately most have been nymphs (such as several from within the QM collection) and therefore they could not be confidently identified as Trolicaphyllium sarrameaense comb. nov. or as Trolicaphyllium erosus comb. nov. nymphs. Unfortunately, in Größer (2008b) the key to species of Chitoniscus sensu lato tried to use the female tegmina radial and media venation pattern to differentiate species, but mixed up the species. Within the key Trolicaphyllium erosus comb. nov. and Trolicaphyllium brachysoma comb. nov. (from New Caledonia) instead key out as the Fijian population and Chitoniscus lobiventris (Blanchard, 1853) and Chitoniscus lobipes Redtenbacher, 1906 (from Fiji) key out as the New Caledonian population. We have reviewed the type specimen photos available on the Phasmid Species Files (Brock et al. 2021; http:// phasmida.speciesfile.org) as well as numerous museum specimens, and always the female tegmina venation allowed accurate distinction of these two genera. Even if you look past this inaccuracy within the key, unfortunately no additional features can be gleaned from the further couplets to allow differentiation of Trolicaphyllium sarrameaense comb. nov. from Trolicaphyllium erosus comb. nov. (as the further couplets discuss abdominal shape, which in these two species is identical/variable). At this moment in time, we still lack significant details about the population of Trolicaphyllium gen. nov. on Grande Terre as material is limited and molecular data has yet to be compared across a wide sampling on the island. With phylliid abdominal shapes sometimes rather variable within a single species, this makes us wonder if Trolicaphyllium sarrameaense comb. nov. is in fact a valid species, or simply a synonym of Trolicaphyllium erosus comb. nov. which was described more than 100 years previously from the same island. Our examination of all type specimens which could be traced has not yet revealed additional features for morphological differentiation besides the overall size of these two species. Hopefully, future molecular analyses from across New Caledonia will reveal if there are several species present on Grande Terre or if it is simply a single species which can range in size from smaller (ca. 40 mm; Trolicaphyllium erosus comb. nov.) to larger (ca. 60 mm; Trolicaphyllium sarrameaense comb. nov.). It is due to this lack of sound molecularly based evidence and the propensity for phylliids to be morphologically variable that we refrain from synonymizing Trolicaphyllium sarrameaense comb. nov. with Trolicaphyllium erosus comb. nov. as we feel a significant decision such as this should be based upon a solid foundation. If future molecular analyses reveal that there is only a single morphologically variable species of Trolicaphyllium gen. nov. on Grande Terre based upon a sampling throughout the island, then we feel a synonymization will be necessary, but not until that time.
The etymology given in the original description is that this name is a toponym, named after the type locality, Sarramea, New Caledonia (Größer 2008a). The original combination was with the masculine genus (Chitoniscus) and therefore in order for the species name to agree in gender with our newly erected genus, the spelling of "sarrameaensis" is changed to the neuter gender "sarrameaense".
Differentiation. Females can be differentiated from Trolicaphyllium brachysoma comb. nov. based upon abdominal shape, as Trolicaphyllium brachysoma comb. nov. are considered to have a spade-shaped abdomen, with smooth margins, versus Trolicaphyllium sarrameaense comb. nov. which has a broad abdominal shape with parallel sides, ending in lobed segments VII and VIII. From Trolicaphyllium erosus comb. nov. the only feature we have been able to identify as useful is the overall length, with Trolicaphyllium erosus comb. nov. ca. 40 mm long versus Trolicaphyllium sarrameaense comb. nov. ca. 60 mm long.
Unfortunately, males of Trolicaphyllium brachysoma comb. nov. and Trolicaphyllium erosus comb. nov. have never been confidently confirmed through breeding or molecular comparison. Based upon the confidently confirmed male/female Trolicaphyllium sarrameaense comb. nov. however, we expect that the male morphology should mirror the female morphology. Most likely the male Trolicaphyllium brachysoma comb. nov. will lack prominent abdominal lobes and the male Trolicaphyllium erosus comb. nov. will have distinct lobes to match with their female counterparts. Based upon the female Trolicaphyllium erosus comb. nov. smaller size, we expect that the male must also be rather small, which could likely be used as a feature for differentiation.
Distribution. To date we have only confirmed adult specimens and observations which are the correct morphology and size of Trolicaphyllium sarrameaense comb. nov. from central Grande Terre (Fig. 21). We have however seen nymph specimens which had characteristically lobed abdomens which may represent this species from other locations on Grande Terre, so we expect this species may be widespread throughout the island. Within the MZPW collection there is a pair of Trolicaphyllium sarrameaense comb. nov. specimens with the data of simply "Lifu", which if true could lend credibility to the hypothesis that perhaps these species are all variable in their abdominal shape (if there is only one species present on Lifou island), but as these are antique and give no other data, we do not take these as highly credible, and therefore exclude this record from further discussion and they are not included within the distribution map (Fig. 21). Or it is possible Lifou island holds several morphologically different species.

Generic identity
With significant confusion surrounding the distribution of Trolicaphyllium gen. nov. species in New Caledonia (due to apparent intraspecific variation and sexual dimorphism), this leaves many records as unidentifiable to species (see Suppl. material 1). Additionally, many records we have seen are of nymphs, which are rather difficult to estimate final adult morphology from (Fig. 30). These many records are still rather interesting as they can help to clarify the distribution of this New Caledonian endemic genus (Fig. 21). For example, a singular record of a male from the AMNH which was collected by Lindsay Macmillan on Maré Island during the Whitney South Sea Expedition and recorded as a "mantis" within the expedition journal, was located within the unidentified Mantodea drawer within the AMNH collection by the first author and represents one of only two records we have seen to date from this island (Macmillan 1938).
With phylliids typically rather island endemic, it is uncertain at this point if the records from Tiga (Fig. 31) and Maré islands are a currently described species (representing a range expansion), or if they are yet to be described species. The latter is possible given that no species have yet been described from these islands. Unfortunately, neither of these records could be sequenced as one was observational and the other too old to sequence, therefore little can be gleaned at this time regarding these islands. With Trolicaphyllium gen. nov. records known from Grande Terre, Lifou, Tiga, Maré, Ile de Bélep, and L'Île-des-Pins islands, we expect that possibly other small islands in New Caledonia such as Ouvéa may also have populations, but to date we are not aware of any observational records or specimens from additional islands (Fig. 21). Hopefully, future sequencing of museum specimens and additional local efforts to sample these smaller islands such as Lifou (which will be necessary as it is the type locality for Trolicaphyllium brachysoma comb. nov.) will reveal the true distribution of Trolicaphyllium gen. nov. species in New Caledonia. Of particular use will likely be immature nymph specimens present within many museum collections which although not very useful for morphology, significantly increase the molecular data if sequenced and can add clarity to species boundaries/intraspecific sequence variation.
Only now are the higher-level relationships within the phylliids being explored, with the most likely sister group to Trolicaphyllium gen. nov. appearing to be the Comptaphyllium (Bank et al. 2021). Geographically speaking, one possibility for "filling in this missing gap" between the New Caledonian clade and the Comptaphyllium could be Pulchriphyllium groesseri (Zompro, 1998), a morphologically unique species known only from the Solomon Islands between New Guinea and New Caledonia. This species is only tentatively placed within the Pulchriphyllium due to the presence of exterior lobes on the tibiae and similar tegmina venation, but due to the lack of fresh material to sequence, this species has not yet been included within any molecular phylogenies. Due to the similar body size and abdominal shape to certain Trolicaphyllium gen. nov. species we wonder if Pulchriphyllium groesseri may indeed be related and could geographically link these two genera, therefore giving more clarity to the relationships within this group.
Additionally, geographically, it is possible that Vanuatu to the north of New Caledonia also has suitability for phylliids and may have been part of their route of colonization to modern day New Caledonia assuming an origin of New Guinea for modern phylliids (Bank et al. 2021). Interestingly however, to date Vanuatu lacks phylliid records which could be used as reference and no museum specimens have been located.
Morphologically it is not uncommon within the phylliids for single species to be rather variable within their abdominal shape, ranging from tapered, smooth and boxy, weakly lobed, to strongly lobed (Hennemann et al. 2009;Cumming et al. 2020b). Review of every museum specimen we could locate has shown a significant variability within the abdominal shape of many nymphs and multiple adults. We hope that future molecular reviews of Trolicaphyllium gen. nov. specimens of variable morphology from throughout Grande Terre and the surrounding islands will reveal how many species there likely are in New Caledonia. Within phylliids there appears to be vastly differing limits geographically for how far a single species will range, with numerous examples of island endemics (e.g., Cumming et al. 2018Cumming et al. , 2019aCumming et al. , 2020a as well as instances of single species spanning extensive geographic ranges (e.g., Cumming et al. 2020aCumming et al. , 2021. It is possible with Trolicaphyllium brachysoma comb. nov. known from Lifou, and the others from Grande Terre, that there are at least two species (at least one from each island) and we hope that future molecular review of a wide sampling of specimens will reveal species geographic boundaries with more clarity. If molecular reviews of numerous specimens from Grande Terre reveal only a single morphologically variable species on Grande Terre, this would likely mean that only two species as presently known would be warranted Trolicaphyllium brachysoma comb. nov. from Lifou and Trolicaphyllium erosus comb. nov. from Grande Terre, with Trolicaphyllium sarrameaense comb. nov. a junior synonym of Trolicaphyllium erosus comb. nov. Hopefully, future molecular reviews will help clarify the Trolicaphyllium gen. nov. diversity within New Caledonia.

Functionally relevant morphological characters
Some morphological features separating Chitoniscus sensu stricto and Trolicaphyllium gen. nov., namely tarsal and egg morphology, can shed further light on the evolutionary history of these groups, as the morphological difference can be a result of adaptations that accompany the process of speciation. Their functional relevance can highlight the functional constraints which led to the specific morphological traits. The overall tarsus morphology in phasmids in general and in Phylliidae in particular is quite conserved, but the differences found are probably results of adaptations to the specific environments of the species (Büscher and Gorb 2017;Büscher et al. 2018a, b). While the male tarsus of both groups is similar and not different in morphology when compared to other Phylliidae , the female tarsi differ in one functional feature. The tarsi of Trolicaphyllium sarrameaense com. nov. females bear an extension in form of a median ridge of the euplantulae on the tarsomeres 2-4 ( Fig. 12B), which is so far not reported for any other phasmid in this form. This accessory ridge can elongate the adhesive surface of the tarsus distally and increase traction along the length of the tarsus, as euplantulae are primarily used for generation of friction in phasmids (Busshardt et al. 2012;Labonte and Federle 2013;Büscher et al. 2020a). However, the arrangement of the euplantulae stabilizes the attachment primarily in the proximal-distal direction, which could be beneficial in the typical postures of these insects in which their legs are arranged circular around the stout body. Furthermore, the elongation of the tarsal chain increases traction on thin stems, if the tarsi grasp around the substrate (Büscher et al. 2020a). A similar effect is additionally achieved by the accessory euplantula on tarsomere 5. This feature is consistently present in all investigated Phylliidae but lacking in many other phasmid species (Vallotto et al. 2016;. The eggs of both groups differ significantly in the modification of their exochorionic surface. While Trolicaphyllium gen. nov. bear distinct microscopic spherical structures on the surface of the eggs (Fig. 17A-D), the eggs of Chitoniscus sensu stricto have a rough, porous surface. These surfaces can be fundamentally different in their function. The spherical structures might be water and dirt repellent, while the rough, porous surface can facilitate water spreading or uptake (Watson et al. 2017). These functions, however, need to be tested in subsequent studies to corroborate their potential benefits.
The presence or absence of pinnae and their respective morphology have been shown to be of taxonomic value for the eggs of Phylliidae already (Clark 1978;Cumming et al. 2020a), this feature, however, is also of striking functional relevance (see. Büscher et al. 2020b, c). The exochorionic pinnae of the eggs of Phyllium philippinicum are involved in a water responding adhesive system that attaches the eggs to different surfaces (Büscher et al. 2020b). Similar fringe-like expansions on their egg shells are reported for other phylliid eggs and hypothesized to be involved in adhesion as well ). Due to their similarity in morphology, it is highly likely that the eggs of Chitoniscus sensu stricto provide the same adhesive capability. These structures probably also carry a glue and expand when they come in contact with water and distribute glue on the corresponding surface. Whether the mushroom-like elements on the surface of Trolicaphyllium gen. nov. eggs provide adhesion and how their potential adhesion responds to water is so far unknown and could be tested in subsequent stud-ies. This could also yield a more elaborate analysis of the evolution of adhesion in the eggs of Phylliidae including closely related lineages with different pinna morphologies.
To conclude, despite more than a century of considering the Fijian and New Caledonian phylliids a single genus, we herein adjust the taxonomy of this polyphyletic clade. Our erection of the Trolicaphyllium gen. nov. reflects the unique aspects of the species of New Caledonia as distinct from their previously considered congenerics (Chitoniscus sensu lato).
Herein we have addressed the several questions presented in the introduction regarding these two groups of leaf insects. Firstly, although these clades are in general similar due to their smaller size and abdominal shapes ranging from smoothly tapered to boxy and strongly lobed, a thorough review of small and microstructures of adults, nymphs, and eggs present a plethora of morphological features for easy and reliable differentiation (Table 1). Secondly, although molecular data for species from these two countries is rather limited, all results have been consistent and in agreement with our morphological findings, with the two clades always corresponding to the geographic distributions (Fiji always sister to the greater leaf insects and New Caledonia a distinct clade nested within the greater phylliids). Thirdly, based upon our review, we assessed what taxonomic option would most clearly correct this issue. With such drastically different morphologies for these two clades, as well as their unique morphologies from all other extant phylliids, our most transpicuous taxonomic path was for clarifying these as genera.
Therefore, based upon years of molecular results, dozens of morphological features, and the significant geographic isolation for these two clades of leaf insect, we feel that the erection of Trolicaphyllium gen. nov. properly addresses the many discrepancies which were glaringly problematic with the Chitoniscus sensu lato.
(United Kingdom) for thoughtful discussion on the Redtenbacher type specimens. Thank you to Christine Rollard (MNHN) who shared notes on Philippe François with us. Thank you to Tony Robillard at (MNHN) for checking the museum logbooks for information regarding the syntype T. erosus females from "Canal Woodia". Thank you to Mark Youles (United Kingdom) and Scott Kirkland for helping to identify the type locality for one of the syntype T. erosus females from "Canal Woodia". Thank you to Tajan Auction House (Paris, France) for sharing a wonderful photograph of Philippe François. Thank you, Amélie Saint-Marc (Campus Condorcet, Paris-Aubervilliers, France) for scanning and sending us selected pages from the archived work within your collection; this was instrumental in learning more about the syntype from Canal Woodin collected by Philippe François. Thank you to René Limoges (Canada) for taking many photographs for this work, as well as for many professional courtesies. Thank you to our scientific illustrator Liz Sisk (USA) who did a beautiful job of illustrating the freshly hatched nymphs for us as no image correctly captures the morphology since nymphs always hold their abdomen or little legs at all angles and seem to be always moving. We thank Michel Saint-Germain, head of collections and research and Maxim Larrivé, director of the Montreal Insectarium for their support of our research.