Systematics of Amphineurus (Rhamphoneurus Alexander) (Diptera: Tipuloidea: Limoniidae)

The subgenus Amphineurus (Ramphoneurus Alexander) (Diptera: Limoniidae), a group of craneflies endemic to southern South America, is revised. The previously described species are redescribed and illustrated, along with fifteen new species: A. (R.) alexanderi sp. nov., A. (R.) amorimi sp. nov., A. (R.) anchoralis sp. nov., A. (R.) anfractus sp. nov., A. (R.) billinghami sp. nov., A. (R.) caleuchus sp. nov., A. (R.) deceptus sp. nov., A. (R.) falcatus sp. nov., A. (R.) immaculatus sp. nov., A. (R.) morphyi sp. nov., A. (R.) oosterbroeki sp. nov., A. (R.) podenasi sp. nov., A. (R.) theischingeri sp. nov., A. (R.) triangularis sp. nov., and A. (R.) stigmaticus sp. nov. In addition, A. (R.) fuscifusus Alexander is considered a nomen dubium. A phylogenetic analysis with a broad taxonomic sampling with all valid species of the subgenus A. (Rhamphoneurus Alexander) and species from all the subgenera of Amphineurus Skuse is performed. The broad outgroup sampling used in the study (including taxa from the entire range of distribution of Amphineurus) suggests that the subgenus A. (Rhamphoneurus) is a monophyletic taxon, that its closest relatives are taxa currently distributed in New Zealand, and that Australasian taxa are paraphyletic with respect to the Neotropical Clade.


Introduction
The southernmost parts of the Southern Hemisphere contain a rich and unique biota of craneflies (Diptera: Tipuloidea) which is poorly understood, but that nevertheless suggests strong biogeographical links between southern South America and Australasia. In a comprehensive review of the subject, Ribeiro and Eterovic (2011) examined in detail the distributional patterns of over 700 species in 30 genera and subgenera of Tipuloidea endemic to areas that constitute what is known in biogeography as The genus Amphineurus includes four subgenera and over 80 species. Amphineurus (Amphineurus), with approximately 42 described species, is distributed in southern South America, New Zealand, New Caledonia, Australia and New Guinea. The Australian species of A. (Amphineurus) were revised by Theischinger (1994Theischinger ( , 1996, but there is no recent taxonomic treatment of the species from the other regions. The subgenera A. (Nesormosia Alexander) (four species) and A. (Nothormosia Alexander) (18 species) are New Zealand endemics. Ale xander (1929a) proposed the subgenus A. (Rhamphoneurus) to include a group of Neotropical species with a long rostrum, which now constitutes a southern South American endemic group (10 species in Chile and Argentina) (Oosterbroek 2022).
It is of historical interest that, when reporting for the first time the occurrence of Amphineurus in the Neotropical Region, Alexander (1929a) regarded this fact as being of great scientific significance. In his own words: "Although the majority of the Chilean species represent a new subgeneric group, one species, A. castroensis […], belongs to the rather extensive aggregation of forms that are allied to A. perdecorus Edwards, and A. pressus Alexander, known hitherto only from New Zealand" (Alexander 1929a: 186). These remarks were made in a context in which Alexander (1929a) defended the hypothesis already raised at that time that Antarctica must have played a role as a land connection between South America and Australasia.
As discussed by Ribeiro and Eterovic (2011), detailed taxonomic and phylogenetic studies of this fauna are of paramount importance for testing the primary hypothesis of biogeographical homology they propose, stating that cranefly taxa now disjunct in different sides of the Pacific were integrated into an ancient biota that underwent considerable level of differentiation before the breakup of Gondwana.
In the present contribution, we provide a detailed taxonomic revision and phylogenetic study of the Neotropical subgenus A. (Rhamphoneurus), with the description of 15 new species and an identification key for all its included taxa. Knowledge of the previously described species of the subgenus A. (Rhamphoneurus) is virtually limited to the original descriptions. In addition, knowledge of the morphology and biology of immature stages of this subgenus remains unknown.
Testing the monophyly of A. (Rhamphoneurus) demanded broad taxonomic sampling including representatives of all subgenera of Amphineurus from the entire distribution range of the genus; i.e., from South America, New Zealand, Australia, Tasmania, New Guinea and New Caledonia.

Material and methods
The specimens of this study belong to the Alexander Collection of Crane Flies of the National Museum of Natural History, Smithsonian Institution, USA (USNM). The holotype of A. (R.) sanus belongs to the Natural History Museum, London, UK (BMNH). C.P. Alexander made preliminary identifications of all examined specimens. Specimens on slides prepared by C.P. Alexander were identified with an asterisk (*). Descriptive terminology follows McAlpine (1981) for most structures, Schneeberg and Beutel (2011) for head morphology, and Ribeiro (2006) for terms of the structures of the male terminalia. The terminology applied to the wing veins follows, with minor changes, the terminology adopted in the "Manual of Afrotropical Diptera" (Cumming and Wood 2017).
When necessary, clarifications of male and female genitalia were made with the method proposed by  using Proteinase K. The genomic material of the clarified specimens was stored. Photographs were taken with a Nikon DS-Ri1 digital camera attached to a Nikon SMZ1000 (stereoscopic) microscope. Deep focus images were stacked with Combine ZP software. Measurements were taken with Carl Zeiss AxioVision (Release 4.8) software. Line drawings were made by camera lucida attached to the microscopes. Described measurements are of the holotypes. Regarding species distribution records, we have included the approximate geographic coordinates of the localities identified on the labels of the studied specimens.
The phylogenetic analysis was made under different criteria: 1) parsimony with equal weights; 2) parsimony with implied weights; 3) Bayesian inference. This was made in order to test the stability of the results under different analytical protocols.
For the parsimony analysis, we used heuristic searches with tree-bisection reconnection (TBR) branch swapping with 1,000 replicates holding up to 100 trees per replication. All characters were treated as unordered. Cases with the character state unknown were scored with '?' and cases where a character is inapplicable with '-'. The analyses were made using TNT (Goloboff and Catalano 2016) and rooted following the outgroup method revised by Nixon and Carpenter (1993). It must be emphasized here that, according to Nixon and Carpenter (1993), a good definition of character polarity within the ingroup does not only depend on the proper definition of the outgroup taxa, but also on the inclusion of characters selected to provide some resolution within the outgroup. In this paper, we have adopted this protocol as much as possible, including an impressive representation of all subgenera of Amphineurus and its morphological variation.
For Bayesian inference, we used the MK model (Lewis 2001). A log normal distribution was used to account for character rate variation. The analysis was made with two simultaneous runs containing 10 million generations of the mcmc chain, each run had four chains. Trees were sampled every 1000 generations with a burning cutoff of 25 percent. Convergence was checked by examining the trace plots using Tracer 1.7 (Rambaut et al. 2018) and the standard deviation of the sample spits. The tracer plots reached stationary and the standard deviation was lesser than 0.01 (0.004). Diagnosis. Amphineurus can be diagnosed by the following combination of characters: gonocoxite bifid; wing with R 2+3+4 aligned with R 2+3 . When there is an element R 4+5 (and therefore, no R 2+3 ), members of the genus are distinguished from other Chioneinae by the closed discal cell.

Taxonomic revision of
Remarks. There is great diversity in the male terminalia of Amphineurus (Figs. 1, 2). There are many variations from the simplified pattern of A. (A.) patya Theischinger (Fig. 1A) to the asymmetric disposition of A. (Nesormosia) (Fig. 2D). Therefore, defining a broad diagnosis for Amphineurus is not so easy. Skuse (1890) defined Amphineurus as a subgenus of Rhypholophus. Skuse's definition fits better to the current subgenus A. (Amphineurus) but does not portray the reality of other species of Amphineurus. His diagnosis included the absence of the discal cell, the radial sector with four branches, "second longitudinal vein angulated", and hind femora one third longer than the intermediate pair. These features are non-exclusive or variable in the genus. Alexander (1920) allocated Amphineurus as a subgenus of Ormosia after highlighting the abundant and conspicuous wing macrotrichia (Fig. 5B). Edwards (1923) redefined the genus to support two New Zealand species. Edwards' diagnosis includes the short mid-leg, short and hairy wing, and radius four-branched. Nevertheless, these features occur in other Chioneinae. Edwards (1923) also referred to the fact that the medial vein forks in M 3+4 and M 1+2 and that the anal vein is sinuous. These features are plastic in the genus. Alexander (1929a) used the definition of Edwards (1923) and emphasized the characteristic "abundant trichiation" of the genus. However, in the same article, Alexander (1929a) described the genus Maietta, which has more wing vestiture than Amphineurus. Some species of the Neotropical subgenus A. (Rhamphoneurus) have less macrotrichia than other subgenera. No more inclusive diagnosis was proposed and recent revisions (e.g., Theischinger 1994Theischinger , 1996 redefined only the diagnosis of the subgenus A. (Amphineurus).
Amphineurus has an 180º rotation of male terminalia (in the words of Edwards: "undergone permanent torsion"). Edwards (1923) also shows that Amphineurus has a long, finger-like gonocoxite process (Fig. 1A). This is evidence of the bifid gonocoxite, a condition present (Fig. 29G) in some Chioneinae with inverted male terminalia. This gonocoxal process forms a protective hood (Kramer 2020) for the appendices of the inverted terminalia.
The Chioneinae genera with inverted terminalia bear great diversity and frequently they have hairy wings similar to Amphineurus. The great amount of wing vestiture resembles Maietta Alexander. The wings of some A. (Amphineurus) (Fig. 32: Node 18) have a distinctive radial sector such as seen in Molophilus Curtis. Some A. (Amphineurus) (Fig. 32: Node 23) resembles Hoplolabis (Parilisia Savchenko). The trifid claspers of A. (Rhamphoneurus) are observed in genera such as Hoplolabis Osten Sacken and Cheilotrichia Rossi. Despite that, in the group of genera with a bifid gonocoxite and inverted terminalia, vein R 2+3 is normally aligned with R 2+3+4 . The exception is a clade of A. (Amphineurus) (Fig. 32: Node 18) that shows resemblances in the radial sector with Molophilus (where R 2+3+4 is lacking), but the former has a closed discal cell among other wing features which places it within Amphineurus.
Description. Habitus of representative species of the subgenus A. (Rhamphoneurus) is shown in Fig. 3. Head as long as wide in lateral view (Fig. 4A). Two ventral cervical sclerites of same proportions, longer than wide, while sclerite in middle of neck nearly horizontal; sclerite near head more vertically orientated (Fig. 4A) 23G); posterior margin may bear medial sclerotized projection (Fig. 28G); darkened in most species (Fig. 20F). Gonocoxite bifid ( Fig. 6) with long ventral branch and short dorsal branch; additional branch of gonocoxite (= mesal lobe) present in majority of species (Fig. 6), very variable, including asymmetry. If asymmetrical, right and left mesal lobes drawn in corresponding position on plates.
Gonostylus bifid ( Fig. 6)  Etymology. The species is named after C.P. Alexander, the remarkable Tipulomorpha expert who described all the species of this subgenus, among many others.
Diagnosis. This species is distinguished by a curved R 2+3 , large markings only near R 2 , r-m longer than basal deflection of R 5 , and R 2+3+4 longer than R 2+3 . Furthermore, the species has characteristic male terminalia with a trapezoidal-shaped spur and symmetrical tip of left mesal lobe. Etymology. The word anchoralis is Latin for "of the anchor". The term refers to the shape of the projection of the male 9th tergite, which resembles an anchor.
Diagnosis. The species is distinguished by the absence of markings near R 1 , and R 2+3 shorter than R 2 . Furthermore, this species has a long vertical spur on the sheath of aedeagus, male tergite IX has a distinctive ridge projection, mesal lobes absent and the posterior margin of the male tergite IX is mostly unpigmented. Diagnosis. This species is distinguished by having vein bM forking into M 2+3+4 and M 1 , absence of markings on R 2 and insertion of r-m desclerotized. Furthermore, the species has distinctive male terminalia with a curved sheath of aedeagus and both mesal lobes similar in length with vertically directed tiny tips.  Etymology. The species is named after Zacariah D. Billingham, a great expert of Australian craneflies.

Description
Diagnosis. This species is characterized by crossvein m-cu uniformly inserted at the fork of bM. Furthermore, this species is characteristized by the male terminalia with a curved sheath of aedeagus, a small right mesal lobe and the most setose ventral branch of gonocoxite. The thorax is also distinguishable by a narrow knob of the halter and katepisternum longer than the meron.

Amphineurus (Rhamphoneurus) caleuchus sp. nov.
h t t p : / / z o o b a n k . o rg / 2 0 7 2 6 C 1 6 -A 3 3 3 -4 F B 5 -A 2 2 6 -BC73116D14FF Etymology. The word caleuchus refers to the Chilean myth of Caleuche, a sentient ghost ship of the Chiloé islands (Febles 2009). The species is named after this myth due to holotype and the majority of specimens being from the Chiloé islands.

Amphineurus (Rhamphoneurus) chiloeanus Alexander, 1969
Amphineurus ( Diagnosis. The species is distinguished by having m-cu strongly sinuous, straight anal vein and beginning of M 1+2 retains sclerotized section. Furthermore, this species has a characteristic sheath of aedeagus with a long basal spur, mesal lobe is very reduced and male tergite IX is remarkable by its rounded posterolateral lobes. Redescription. Wing length 7.37 mm, width 2.14 mm.  15G): Male tergite IX faintly darkened at posterior margin, with wide, shallow U-shaped median notch; posterolateral lobes rounded. Dorsal branch of gonocoxite poorly developed, shorter than half-length of ventral branch. Lobe of gonostylus with elongate lobule, longer than wide, longer than half-length of lobe of gonostylus, but narrower than stem. Clasper with differently shaped branches: lateral branch straight and long; medial branch curved, wide club-shaped. Mesal lobes atrophied, similar in shape to dorsal branch of gonocoxite. Sheath of aedeagus long and sinuose, darkened almost entire length, with pointed perpendicular spur projecting at end of straight section.
Remarks. Some specimens of this species were misidentified by C.P. Alexander as A. (R.) nothofagetorum Alexander. This species resembles A. (R.) glabristylatus Alexander but differs mainly in attachment of m-cu, length of cercus, shape of male tergite IX, sheath of aedeagus and mesal lobe.

Amphineurus (Rhamphoneurus) extraordinarius Alexander
Amphineurus ( Remarks. The species Amphineurus extraordinarius was originally included in the subgenus Amphineurus. In all the phylogenetic analysis we performed (see below in this paper), this species is the sister-group of the other species of Rhamphoneurus, and we think it should be transferred to this subgenus. This new subgenus placement is therefore based on the phylogenetic analysis, but specimens of A. extraordinarius were not available when the analysis was finalized. Consequently, a redescription of the species is not possible at this time.

Amphineurus (Rhamphoneurus) deceptus sp. nov.
h t tp : //z o o b a n k . o rg / C E7 F 2 5 3 C -D C 8 9 -4 D A C -9 5 B 7 -1C4D89095BB9 Etymology. The word deceptus is Latin for "deceived". The term refers to the fact that specimens of this new species were misidentified by Alexander as two different species.
Diagnosis. This species is distinguished by the darkened distal quarter of the wing, and bM is uniform. Furthermore, the species is characterized by the distinctive male terminalia with clasper of gonostylus with setae on medial branch and straightened base of sheath of aedeagus. Etymology. The word falcatus is Latin for "armed with scythes". This name is in reference to the sheath of aedeagus, which resembles a scythe.
Diagnosis. This species is distinguished by absence of markings near R 2 and R 2+3+4 longer than R 2+3 . Furthermore, this species has distinctive male terminalia with a curved sheath of aedeagus, only the right mesal lobe folded, and the longest lobule of the subgenus, as long as the remainder of the lobe of gonostylus. Remarks. Amphineurus (R.) fuscifusus Alexander is the only species of the genus in which the male is unknown. Additionally, the only identified specimen is the female holotype, currently housed in the Natural History Museum (BMNH). The absence of an associated male specimen and its terminalia is problematic because this is an important source of characters and taxonomic data. Lacking a well-defined diagnosis in the description, the erection of the new species was proposed mostly due to subtle differences in coloration. The only considerable difference in coloration in A. (R.) fuscifusus is the pale halter, which is unlike the yellowish brown halter of A. (R.) nothofagetorum Alexander.
Some morphological data described by Alexander (1929a) suggest this specimen shares similar features with clade 33 (Fig. 32). The presence of long setae on the antenna, "weak branches" in medial veins, three stripes on the scutum, and a marking near the fork of Rs are features shared between the holotype and this group. The holotype was collected in the same location as the holotype of A. (R.) nothofagetorum Alexander. However, the absence of males added to the fact that there is only a single known specimen makes the synonymization or identity confirmation of the species problematic. For these reasons, we define Amphineurus (Rhamphoneurus) fuscifusus Alexander as nomen dubium.

Amphineurus (Rhamphoneurus)
glabristylatus Alexander, 1929a Amphineurus ( Diagnosis. This species is distinguished by the sinuous anal vein and straight m-cu. Furthermore, this species has characteristic male terminalia with a folded medial branch of the clasper of gonostylus without setae, an unpigmented male tergite IX, sinuous sheath of aedeagus, and the right mesal lobe is longer than the left mesal lobe, with a cylindrical apical tip.  18F). Anatergite taller than katatergite; prescutum with long setae (Fig. 29B); mediotergite with notch on posterior margin (Fig. 18F). Halter with dilated knob. Wing (Fig. 18D) with several markings, including subtle mark-ing on A 1 ; slightly darkened areas along medial veins; markings on Rs, R 2+3+4 and m-cu; strong markings on R 2 and near end of R 1 ; veins near fork of bM (Fig. 18C)  Etymology. The word immaculatus is Latin for "unstained". The term refers to the fact that the wing has no pterostigma or markings.

Amphineurus
Male tergite IX faintly darkened at posterior margin, with wide V-shaped median notch; posterolateral lobes nearly triangular; ventral branch slender, with concentration of setae on distal tip. Dorsal branch of gonocoxite shorter than half length of ventral branch. Lobe of gonostylus with heart-shaped lobule, more than twice as wide as lobe of gonostylus. Clasper of gonostylus with differently shaped branches: lateral branch straight; medial branch curved, oblong-shaped. Mesal lobes of gonocoxite symmetrical, both with narrow, elongate, straight extremity. Sheath of aedeagus straight, slightly bent near tip, darkened at distal one-third.
Remarks. Some specimens of this species were misidentified by C.P. Alexander as A. (R.) insanus Alexander. This species resembles A. (R.) glabristylatus Alexander but differs mainly in the attachment of m-cu, wing markings, shape of male tergite IX and mesal lobe.

Amphineurus (Rhamphoneurus) sanus Alexander, 1929a
Amphineurus ( Diagnosis. This species is distinguished by a subtle spot on R 2 . Furthermore, this species has a remarkable sheath of aedeagus with a long horizontal spur, male tergite IX with a distinctive rectangular projection, mesal lobes absent and posterior margin of male tergite IX pigmented.  Fig. 28A): Rostrum bifurcated with short appendices; first palpal segment shorter than terminal segment, terminal palpal segment dilated; scape narrower than pedicel; terminal flagellomere shorter and thicker than penultimate; antenna with several long setae.
-Male terminalia (Fig. 28G): Male tergite IX darkened on posterior margin, with wide median U-shaped notch; rectangular projection in two points on either side of extension; posterior margin with projection flat. Dorsal branch of gonocoxite very reduced. Lobe of gonostylus with pointed lobule shaped similar to folded leaf. Clasper of gonostylus with differently shaped branches: lateral branch long, bent around middle; medial branch curved into short stub. Mesal lobes absent. Sheath of aedeagus dark along almost entire length, angulated with smooth curvature, with setae on distal portion of lateral margin, as well as large distal horizontal lateral projection.
Remarks. Some specimens were misidentified by C.P. Alexander as A. (R.) nullus Alexander. The rostrum of the holotype is longer than in the other species of this subgenus. This species resembles A. (R.) anchoralis sp. nov. but differs mainly in M 1+2 , shape of projection on male tergite IX, ventral branch of gonocoxite and spur of sheath of aedeagus.

Amphineurus (Rhamphoneurus) stigmaticus sp. nov.
Etymology. The word stigmaticus is Latin for "marked". The term refers to the darkened mark near the tip of the wing of this new species.

Remarks
Diagnosis. This species is distinguished by an absence of markings near R 2 , r-m as long as basal deflection of R 5 , and cross-vein m-cu straight and inserted at bM. Furthermore, this species is characterized by male terminalia with curved sheath of aedeagus, straight medial branch of clasper of gonostylus and mesal lobes with slender, vertical tips. The thorax is also distinguishable by narrow knob of halter and katepisternum as long as meron.  30F): Anatergite shorter than katatergite. Halter with narrow knob. Wing (Fig. 30D) clear of distinctive markings; veins near fork of bM (Fig. 30C) visible; M 1+2 longer than basal deflection of M 1 ; R 2 faded; R 2+3+4 longer than R 2+3 . -Female terminalia (Fig. 30H): Female tergite IX shorter and wider than tergite X; hypogynial valve very slender; cercus longer than hypogynial valve, thick, and slightly curved dorsad. -Male terminalia (Fig. 30G): Male tergite IX with V-shaped median notch. Distal tip of ventral branch of gonocoxite expanded, setose, with slight curve ventrally. Dorsal branch of gonocoxite shorter than half-length of ventral branch. Lobe of gonostylus with rounded lobule; lobule longer than wide, more than half as long as lobe of gonostylus, wider than stem. Clasper of gonostylus with similarly shaped branches: lateral branch rounded; medial branch conical. Mesal lobes of gonocoxite asymmetrical; both with long, narrow, straight distal projection; right mesal lobe wider. Sheath of the aedeagus curved halfway along length, darkened only on tip.

List of characters
Here the characters and their states are defined (the number of steps of each character as they were optimized in the chosen tree is shown within square brackets). In some cases, the understanding of the morphological features under description requires examination of the accompanying illustrations. The data matrix, with a total of 54 terminal taxa and 112 characters, is shown in Supplementary Material 1.
Head 0 Size of rostrum: (0) shorter than remainder of the head; (1) longer or equal to remainder of the head.

Phylogenetic results
The parsimony analysis with equal weights resulted in one most parsimonious tree of 209 steps, Consistency Index (CI) = 0.58 and Retention Index (RI) = 0.86. The analyses with implied weighting (k ranging 1-6) resulted in only one most parsimonious tree with the same topology, shown in Figs. 32 and 33. The Bayesian analysis provided a tree very similar to that of the parsimony analysis, and it is depicted here in Fig. 34. The main difference is a less resolved tree in the Bayesian analysis, especially regarding the relationships within the outgroups. Within the ingroup, there are minimal topological differences; i.e., the collapse of a single node and the position of one species.
To discuss the distribution of the characters in the tree, we use the result of the analysis with implied weighting (with k=3), focusing on the more relevant characters. We believe that our extended outgroup sampling has an enormous heuristic value for future research on the genus Amphineurus, and although we will not discuss here the synapomorphies of the clades recovered in the outgroup, we have numbered all the nodes of the cladogram for future reference. In the lines below, we discuss the synapomorphies and the relationships recovered for the subgenus A. (Rhamphoneurus) (Node 25, Fig. 33).
The synapomorphies of each clade (as provided by the analysis with implied weighting with k=3) are listed within square brackets in the discussion of each node in a standardized way: [number of character: plesiomorphic state -synapomorphic state]. Homoplastic apomorphic states are indicated by an asterisk (*). The clade is supported by the bifurcated lobe of gonostylus [77: 1]. This feature is present also in clade 9 (outgroup), but there are differences. In A. (Rhamphoneurus), the bifurcation of the lobe is a lobule [78 : 1]. Besides, the position of crossvein m-cu [50: 1] reinforces this clade.

Node 25 -subgenus
In our analysis, the species A. extraordinarius is the sister group of all the other species of A. (Rhamphoneurus). Alexander (1939) commented that this Chilean species is very different from the rest of the Neotropical fauna, justifying the choice of the species epithet (meaning "extraordinary"). Many features are discordant with the other Neotropical Amphineurus. Its rostrum is short, tergite IX is trilobate, the apex of the sheath of aedeagus is serrated and the verticils of the antennae are unilaterally distributed.
Amphineurus extraordinarius has M 3+4 bifurcated, as in the Neotropical A. castreoensis. However, this resemblance in the medial sector is a symplesiomorphy. A considerable difference between them is the open discal cell of A. extraordinarius as opposed to a closed one in A. castroensis. The very short M 3+4 of A. extraordinarius is a necessary step to the probable capture of M 3 in Node 26. In our view, this species is a representative of a short-rostrum lineage of A. (Rhamphoneurus), and we expand the diagnosis of the subgenus to include it. The terminalia in this clade is very different from that of other Amphineurus. The distal half of the male tergite IX is dark [55: 0] (Fig. 17G). The sheath of aedeagus is curved [100: 2], with a single marking near to its apex [98: 0] (Fig. 17G). However, the shape of the sheath of aedeagus changes sometimes in the group. Another relevant feature in this clade is the mesal lobe. The mesal lobe of Node 26 has a dark end [87: 1], in contrast with other mesal lobes. Alexander (1925) cited the mesal lobe of A. pulchripes as "abruptly paler" compared with that of A. (Rhamphoneurus) species. Furthermore, the mesal lobe in Node 26 has typical lateral rows of setae [88: 0] (Fig. 20F). The shape of the mesal lobes varies considerably in derived clades in this group, even disappearing in Node 42.  (Fig. 13G). Wings of this clade have a notable absence of markings, even the pterostigma [31: 0]. The maximum of this reduction is in A. (R.) immaculatus sp. nov. (Fig. 19G). Also, the species of this group show a characteristic clasper of gonostylus (Fig. 13G), with branches of the same length [73: 0]. Despite the shared similarities in the wings and gonostylus, the mesal lobes vary in this group. These species show one tip of the mesal lobe straight and the other bent (Fig. 17G). In addition, the node is supported by the shorter prescutum [9: 1] (Fig. 17F) and by characteristic lateral dark lines in the dorsal scutum [11: 1].
Node 29. Synapomorphies: [10: 1-0*]; [50: 1-0*]. This clade is supported by a simple scutum, without dorsal markings [10: 0]. Also, these species do not show other thoracic markings, an uncommon condition for this subgenus. This clade has species with translucent wings with great concentration of setae (Fig. 13G). Furthermore, the insertion of crossvein m-cu distal to the fork of Rs [50: 0] is another synapomorphy of this node (Fig. 13G) that is only observed in the outgroup as a homoplasy. Amphineurus (R.) theischingeri sp. nov. and A. (A.) billinghami sp. nov. are species with high morphological similarity. These species have very similar wings and male terminalia, although the length of the mesal lobes is crucially different. The node is supported by the labrum shorter or equal to half the size of the hypopharynx [3: 1] (Fig. 13B). Also, the wings of this clade have the vein R 5 as dark as the other veins [35: 1] (Fig. 13D).
Other synapomorphies are seen in the thorax: notably long setae on the anteropronotum [14: 0], the anterodorsal margin of the anepisternum above the posterodorsal margin [18 : 1], and the posterior basalare sclerite more than 3× longer than wide [19: 1] (Fig. 13E). The knobs of the halteres in both species also have the characteristic shape of being more than 2.5× longer than wide [24: 1].  (Fig. 16C). There are not only faded veins, but also a lack of pigmentation in their vicinity, giving this section a lighter appearance. Alexander (1929a) points out "weak branches" near to the fork of bM. It may be related to the thyridium, a desclerotization of the medial vein before the fork (Shcherbakov et al. 1995). In some species, this desclerotization of the nodal flexion line extends beyond the fork and affects crossvein m-cu and vein M 1+2 .
Despite the co-occurrence, these faded sections are independent events in unrelated veins such as the tip of vein R 1 [36: 1], vein M 1+2 [39: 1], the insertion of m-cu [41 : 1], and the distal section of r-m [38 : 1]. Other evidence to the independence of these evolutive novelties is the reversion of some of them independently. The descle- rotization of the tip of R 1 is reversed in Node 44. These transformations are common in the subgenus. In Node 27, independently, there are species with vein R 2 faded.
In addition, this clade is characterized by a distinct clasper of gonostylus (Fig. 16G), with the medial branch shorter than the lateral branch [74: 1] It is reversed only in A. (R.) apiculatus (Fig. 12F). Besides that, the dorsal scutum has lateral dark lines [11: 1]. These lines are present also in Node 28 and reversed in Node 40. The fused anepisternum is also seen in A. (R.) falcatus sp. nov. (Fig. 17E). Despite the resemblances, the male terminalia of both species are different. The main difference is a reversion of Ch. 74 in A. (R.) apiculatus (Fig. 12F). In this species, the claspers have branches of the same length [74: 0].  This node groups A. (R.) insanus and A. (R.) alexanderi sp. nov. These species are so similar that all specimens of A. (R.) alexanderi sp. nov. were misidentified by C.P. Alexander as A. (R.) insanus. The defining synapomorphy of this clade is the right mesal lobe with symmetrical sides in the base [93: 1] (Fig. 6).

Biology
The label data of the Amphineurus specimens suggest a univoltine development. The adults are found from the end of November to April, but the majority of species occur from December to January. There is a wide range of tolerated altitude, from sea-level to 1600 m. Most specimens and species were collected in humid environments. Amphineurus (A.) submolophilinus is suggested to be found near streams (Alexander 1923). The larvae of A. (A.) hudsoni have been found in wet decayed fallen leaves of Nikau (Rhopalostylis sapida Wendl and Drude), an endemic New Zealand palm (Edwards 1923). Amphineurus are frequently associated with plants. Amphineurus (A.) bickeli are found in montane forest (Theischinger 1996), and A. (A.) molophilinus in scrub (Alexander 1922a). Amphineurus (A.) breviclavus Alexander, 1924 and A. (N.) otagensis Alexander, 1922a are found among the undergrowth ferns in beech forests. The lower vegetation in beech forests is the environment where A. (A.) campbelli Alexander, 1922b and A. (N.) nothofagi Alexander, 1925 also are found.
In South America, data is more scarce. However, A. extraordinarius is associated with trees of Araucaria Kuntze (Alexander 1939). In New Zealand, A. (N.) longi and A. (A.) kingi are collected near kauri trees (Agathis australis Loudon), an endemic Araucariaceae (Alexander 1950). As evidenced by a recent revision of long-proboscid limoniids (Oosterbroek and Lukashevich 2021), Nothofagus forests bear genera with long rostrums but without direct evidence of pollination in angiosperms. There is an evolutionary scenario of long rostrum craneflies associated with ancient gymnosperm forests that may have eventually adapted to angiosperms. This change has occurred independently in many lineages (Ribeiro 2008;Oosterbroek and Lukashevich 2021). Alexander (1929b) noticed some Amphineurus species in grass steppe and low scrubs of upper montane regions. The most direct association between Amphineurus and angiosperms is with A. (A.) operculatus, found in shrubs of dogwood (Cassinia aculeata Labill.: family Asteraceae) (Alexander 1924). Some Amphineurus (mainly the subgenera A. (Amphineurus) and A. (Nothormosia)) have spotted wings and striped legs. Pritchard (1983) observed that cranefly species with patterned wings are generally found in open woodland. According to Edwards (1923), some Amphineurus species may be mimetic of spiders, as in A. (A.) perdecorus, when resting, assumes a "decidedly spider-like appearance". It may suggest a possible mimetic behavior of this species.

Biogeography
The biogeography of the genus Amphineurus is an old topic. Alexander (1929a) identified Neotropical species of the genus and argued that the presence of the genus in both South America and New Zealand were evidence of past Antarctic land connections. The biogeographical affinities of the cranefly fauna of southern South America and Australasia is reinforced by Ribeiro and Eterovic (2011) and .
Although the focus of our study is the subgenus Amphineurus (Rhamphoneurus) (Fig. 35), we used a broad outgroup sampling. These outgroup species embrace the entire range of distribution of Amphineurus. Based on the outgroup method (Nixon and Carpenter 1993), our analysis is a step forward in resolving the relationships within the outgroups. The outgroup relationships indicated, with some confidence, that the closest relatives of A. (Rhamphoneurus) are in New Zealand. Ribeiro and Eterovic (2011) criticized the simplistic labeling of the austral distributions of Tipuloidea within the South Pacific Track as Gondwanan, in the sense that the final breakup of Gondwana may have not been the event responsible for the differentiation of the lineages. If the final breakup of Gondwana was the reason behind the diversification of the taxa, we should expect monophyletic groups in each of the resultant landmasses.
Concerning taxa distributed in Australasia and South America, there is a recurrent asymmetry, also not expected by vicariance driven by the fragmentation of Gondwana, in which Australasian taxa (notably, New Zealand taxa) constitute a grade (i.e., are paraphyletic) relative to South American taxa. This asymmetry has been found, for instance, in phylogenetic studies of other components of this Trans-Pacific Biota (such as the genus Aphrophila, recently revised by ) where the New Zealand taxa correspond to a grade, with some Zealandia (the continental landmass sensu Mortimer et al. 2017) endemics more closely related to South American groups. Once again, as in the case of Amphineurus dealt with in this study, we have observed the same pattern.
The asymmetry just mentioned, and other general distributional patterns non-congruent with the sequence of breakup of Gondwana can be interpreted as evidence of a heterogeneous set of ancient, pre-rift distribution ranges in the area. Recent phylogenetic studies including both fossil and recent representatives of Tipuloidea (Lukaschevich and Ribeiro 2019) indicate that the main Tipuloidea lineages (family level clades) have differentiated, at least, from the Triassic (stem lineages) to Upper Jurassic times (more derived families). During this time interval, a lot of pre-rift differentiation must have occurred before the final continental divisions. The separation of continental landmasses is only the final result of long-lasting tectonic events, which caused a lot of in situ differentiation by altering the local landscapes.
One alternative explanation could be by a posteriori dispersal from Australasia to South America, but the poor dispersal capabilities of some (if not all) of the groups in question and their very particular habitat needs (for instance, larvae of Aphrophila craneflies are restricted to fast-flowing waters in mountainous areas) does not support such views. In addition, assuming independent trans-oceanic dispersal seems odd with a unifying view of biogeography as the result of an evolving geological setting (Nelson and Platnick 1981;Heads 2015), and, unfortunately, such ideas became so uncritically accepted in recent times, linked with naïve interpretations of minimum ages of clades as provided by molecular clock estimates (Heads 2015).
In the words of Heads (2013: 85): "If a group is in Australia, New Zealand and South America, but the groups in each area are not monophyletic, as in craneflies […], the phylogeny cannot have been caused simply by the rifting of a homogeneous population. Instead, the pattern suggests that differentiation developed in a long phase of intracontinental rifting that was a precursor to Gondwana breakup with seafloor spreading. The whole breakup process was drawn out and included early phases of "failed" rifting; […] This suggests that the break seen in craneflies […] -between New Zealand + South American groups, and Australia + South American groups -was associated with Jurassic-Cretaceous extension in the West Antarctic Rift System, not seafloor spreading that later broke Gondwana." Lukashevich and Ribeiro (2018) estimated that Tipuloidea must have existed at least since the Middle Triassic (ca. 242-247 m.y.a.), and therefore, the group is old enough to have undergone major pre-rift differentiation and cladogenesis. The alternative hypothesis, i.e., assuming post-rift trans-oceanic dispersal, is not only at odds with the biology and ecological constraints of the groups in question, but with an entire biogeographic research agenda that seeks to understand biogeographic patterns in terms of large integrated processes capable of explaining a plethora of biological phenomena under a unifying view, i.e., that the Earth and life evolve together.

Acknowledgments
We are very grateful to Dr. Wayne Mathis, for the loan of the specimens used in this study (as well as for his unconditional support to the work of GCR for many years) when he was still working as curator of Diptera at the NMNH, Washington. We also thank one anonymous reviewer, grant 2010/00557-7 and 2020/02844-5 to GCR).

References
Alexander CP (