Descriptive terminology follows Cumming and Wood (2017). Wing vein nomenclature follows the interpretation of de Jong (2017), while the homology of male terminalia follows Ribeiro (2006). Abbreviations used in the figures are as follows: A, anal vein; aed, aedeagus; b, bulge of gonocoxite; bM, basal medial vein; cg, clasper of gonostylus; CuA, anterior cubital vein; CuP, posterior cubital vein; cv, supernumerary cross-vein; d, discal cell; gx, gonocoxite; h, humeral vein; ib, interbases of the aedeagal complex; lg, lobe of gonostylus; M, medial vein; pm, parameres; R, radial vein; Rs, radial sector veins; Sc, subcostal vein; t9, ninth abdominal tergite.

Specimens examined for the comparative study are listed in Table S1. We thoroughly reviewed all taxonomic descriptions available in the literature regarding the Idioptera-Eloeophila complex. Many of the descriptions used are very informative, at least for the main morphological wing features. The taxonomic literature of the group includes some recent revisions and has a substantial collection of photos and recent illustrations, which provide a wealth of information on the morphology of the species, which could be successfully used in our comparative study.

We analyzed all available images, revisions and descriptions. While we directly observed only a few specimens, our examination encompassed every published information within these genera. We selected at least one species of each morphological unit taking in mind the diversity of places, number of available material, problematic and contested taxa, and revised species. Our sampling was a balanced expression of the morphological diversity avoiding redundancy (observed previous and posterior the phylogenetic analysis). Guided by this comprehensive review, we selected species that serve as representatives of the spectrum of variation.

As discussed posteriorly, most species, especially those related to or grouped around E. maculata (Meigen, 1804), exhibit a relatively homogeneous morphology. Including a large number of these similar species in our taxonomic sample would result only in redundant information. Therefore, we directed our focus towards more diverse and heterogeneous groups of species. We are confident that our data matrix provides a comprehensive representation of the morphological variation within our study group.

Outgroup taxa for this study consisted of species from the genera Eupilaria Alexander, 1932, Phylidorea Bigot, 1854 and Euphylidorea Alexander, 1972, with the root placed at Eupilaria phoenosoma (Alexander, 1931). The selection of outgroups was based on a previous study by Ribeiro (2008), which included a substantial sample of Limnophilinae lineages and genera with affinities with Eloeophila.

The phylogenetic analysis was conducted based on the principle of parsimony. For character scoring, missing scoring of a character for a taxon was entered as ‘?’ and inapplicability of a character to a taxon as ‘–’. The most parsimonious trees were obtained using both Equal Weights (EW) and Implied Weighting (IW) methods (Goloboff 1993). IW analysis involved combining various values of k (ranging from 2 to 6). Heuristic searches were conducted using tree-bisection reconnection (TBR) branch swapping, with 10,000 replicates and holding up to 100 trees for each replication. All characters were treated as unordered. The analyses were performed using TNT (Goloboff et al. 2016), and the rooting was conducted following the outgroup method as revised by Nixon and Carpenter (1993). The characters used are defined and listed in the following (the number of steps for each character, for k=3, is shown within square brackets). The data matrix with 25 terminal taxa and 73 characters is presented in Table S2.

0. Length of antenna: 0 – shorter than combined length of head and thorax; 1 – longer than combined length of head and thorax. [4]

1. Shape of wing: 0 – middle of wing straight; 1 – middle of wing dilated (Fig. 6b). [1]

2. Wing, spur at base of Rs: 0 – absent; 1 – present (Fig. 5a). [5]

3. Wing, Rs inclination: 0 – straight or slightly inclined; 1 – inclined downwards (Fig. 4a). [1]

4. Wing, tip of R₁: 0 – straight; 1 – curved (Fig. 4b). [1]

5. Wing, length of distal section of R₁ (distal to point of contact with R₂): 0 – shorter than twice length of R₂; 1 – longer than twice length of R₂. [1]

6. Wing, length of R₂₊₃₊₄: 0 – longer than R₂₊₃; 1 – shorter than or equal to R₂₊₃. [3]

7. Wing, direction of tips of R₃ and R₄: 0 – parallel; 1 – divergent. [1]

8. Wing, shape of R₄: 0 – straight; 1 – curved. [1]

9. Wing, M₁₊₂: 0 – unbranched; 1 – branched. [1]

10. Wing, length of section of M₁₊₂ distal to point of connection with m-m (distal M₁₊₂): 0 – shorter than twice length of M₂; 1 – longer than twice length of M₂. [1]

11. Wing, tip of M₂: 0 – straight; 1 – curved downwards. [1]

12. Wing, tip of M₄: 0 – curved downwards; 1 – straight. [2]

13. Wing, length of discal cell: 0 – at least 3 times longer than wide; 1 – shorter than 3 times longer than wide. [3]

14. Wing, length of m-m: 0 – shorter than half length of basal deflection of M₃; 1 – longer than half length of basal deflection of M₃. [5]

15. Wing, distance between tips of M₂ and M₃: (0) shorter than distance between tips of M₃ and M₄; (1) longer than distance between tips of M₃ and M₄. [1]

16. Wing, supernumerary m-cu: 0 – absent; 1 – present. [1]

17. Wing, insertion of supernumerary m-cu: 0 – distal to level of origin of Rs; 1 – proximal to level of origin of Rs. [1]

18. Wing, tip of anal vein: 0 – straight or slightly arched; 1 – strongly arched (Fig. 5b). [1]

19. Wing, spur in tip of A: 0 – absent; 1 – present (Fig. 6b). [1]

20. Wing, origin of C thinning: 0 – between veins R₃ and R₄; 1 – between veins R₄ and R₅. [1]

21. Wing, marking on base of wing: 0 – absent; 1 – present (Fig. 5a). [1]

22. Wing, marking around sc-r: 0 – absent; 1 – present (Fig. 3a). [1]

23. Wing, marking on tip of Sc: 0 – absent; 1 – present (Fig. 5b). [1]

24. Wing, marking around humeral vein: 0 – absent; 1 – present (Fig. 3a). [1]

25. Wing, marking on distal margin of arculus: 0 – absent; 1 – present (Fig. 4a). [1]

26. Wing, marking between humeral vein and origin of Rs: 0 – absent; 1 – present (Fig. 4a). [1]

27. Wing, marking on base of Rs: 0 – absent; 1 – present (Fig. 2b). [1]

28. Wing, marking on tip of R₃: 0 – absent; 1 – present (Fig. 2b). [1]

29. Wing, marking on tip of R₄: 0 – absent; 1 – present (Fig. 4a). [1]

30. Wing, marking around basal deflection of R₅: 0 – absent; 1 – present (Fig. 2b). [1]

31. Wing, size of marking around basal deflection of R₅: 0 – not surpassing r-m; 1 – reaching fork of bM (Fig. 5b). [1]

32. Wing, marking on tip of R₅: 0 – absent; 1 – present (Fig. 6b). [1]

33. Wing, marking on tips of medial and cubital veins: 0 – absent; 1 – present (Fig. 5b). [1]

34. Wing, marking on distal side of discal cell: 0 – absent; 1 – present (Fig. 5b). [1]

35. Wing, marking around m-cu: 0 – absent; 1 – present (Fig. 2b). [1]

36. Wing, connection of marking around m-cu with marking around medial fork and r-m: 0 – not connected; 1 – connected (Fig. 6b). [1]

37. Wing, marking on base of CuA: 0 – absent; 1 – present (Fig. 5b). [1]

38. Wing, double marking on anal vein: 0 – absent; 1 – present (Fig. 5a). [1]

39. Wing, distance between two anal vein markings: 0 – close to each other, i.e. on same part of anal vein (Fig. 5a); 1 – far from each other, i.e. on different parts of anal vein such as one at base and one near end (Fig. 6a); 2 – fused (Fig. 5b). [1]

40. Wing, marking on anal lobe: 0 – absent; 1 – present (Fig. 6b). [2]

41. Wing, patch connecting pterostigma with marking around basal deflection of R₅: 0 – absent; 1 – present (Fig. 6a). [1]

42. Wing, patch connecting marking around sc-r with marking at basal deflection of R₅: 0 – absent; 1 – present (Fig. 6b). [1]

43. Wing, patch connecting marking around m-cu with marking at tip of CuA: 0 – absent; 1 – present (Fig. 6a). [2]

44. Wing, patch connecting marking around m-cu with marking at tip of CuP: 0 – absent; 1 – present (Fig. 6a). [1]

45. Wing, patch connecting marking around humeral vein with marking at base of CuA: 0 – absent; 1 – present (Fig. 6b). [1]

46. Wing, color of patches: 0 – uniformly dark; 1 – with a light core and a dark frame (Fig. 6a). [1]

47. Wing, spots: 0 – absent; 1 – present (Fig. 6b). [1]

48. Bulge in base of gonocoxite: 0 – absent; 1 – present (Fig. 7b). [1]

49. Border of margin of bulge in base of gonocoxite: 0 – smooth; 1 – serrated (Fig. 7b). [1]

50. Distal bulge in gonocoxite: 0 – absent; 1 – present (Fig. 7d). [2]

51. Shape of clasper of gonostylus: 0 – straight; 1 – curved (Fig. 9h). [1]

52. Width of clasper of gonostylus: 0 – flattened (Fig. 9d); 1 – rounded and thick (Fig. 9c). [1]

53. Dorsal hump on base of clasper of gonostylus: 0 – absent; 1 – present (Fig. 9e). [1]

54. Ventral margin of clasper of gonostylus: 0 – straight (Fig. 9b); 1 – pointed medially (Fig. 9j). [1]

55. Distal end of clasper of gonostylus: 0 – bifurcated in two teeth (Fig. 9c); 1 – non-bifurcated (Fig. 9h). [3]

56. Distance between teeth of clasper of gonostylus: 0 – close to each other (Fig. 9j); 1 – widely separated (Fig. 9c). [1]

57. Length of ventral tooth of clasper of gonostylus: 0 – shorter than or similar to dorsal tooth (Fig. 9j); 1 – much longer than dorsal tooth (Fig. 9c). [1]

58. Margin between teeth of clasper of gonostylus: 0 – smooth (Fig. 9g); 1 – serrated (Fig. 9c). [1]

59. Anterior margin of clasper of gonostylus: 0 – smooth (Fig. 9a); 1 – serrated (Fig. 9b). [4]

60. Crest on anterior margin of clasper of gonostylus: 0 – absent; 1 – present (Fig. 9h). [1]

61. Pointed spur on crest of gonostylus: 0 – absent; 1 – present (Fig. 9e). [1]

62. Orientation of dorsal tooth of clasper of gonostylus: 0 – straight (Fig. 9a); 1 – bent (Fig. 9i). [1]

63. Connection of interbases of the aedeagal complex: 0 – fused (Fig. 6a); 1 – non-fused (Fig. 7d). [2]

64. Point of contact between interbases: 0 – angulated (Fig. 7a); 1 – straight (Fig. 7d); 2 – curved (Fig. 8a). [2]

65. Tip of distal branch of interbase: 0 – straight (Fig. 7b); 1 – bent or curved (Fig. 8b). [3]

66. Orientation of distal branch of interbases: 0 – inclined (Fig. 7c); 1 – straight, i.e., parallel to aedeagus (Fig. 8c). [2]

67. Ventral apodeme of parameres: 0 – fused with sheath of aedeagus (Fig. 7a); 1 – detached from sheath of aedeagus (Fig. 7c). [1]

68. Orientation of ventral apodeme of parameres: 0 – pointing laterally (Fig. 7b); 1 – pointing medially (Fig. 8b); 2 – pointing anteriorly. [2]

69. Long lateral branches of sheath of aedeagus: 0 – absent; 1 – present. [1]

70. Lower shelf (sheath of aedeagus): 0 – not developed; 1 – developed (Fig. 8b). [1]

71. Humps of lower shelf: 0 – absent; 1 – present (Fig. 8d). [1]

72. Median projection of lower shelf: 0 – absent; 1 – present (Fig. 8d). [1]

In the phylogeny of Limnophilinae conducted by Ribeiro (2008), the clade Eloeophila + Idioptera was found to be related to Eupilaria and Phylidorea, primarily because of their blade-like interbases (Fig. 7). Other studies have shown that these genera are related (Edwards 1921; Podeniene 2004). Oosterbroek and Theowald (1991) found that Euphylidorea may be related to Phylidorea (Fig. 2b) and Eloeophila. However, the study of Ribeiro (2008) did not support this relationship. Despite this, these genera belong to a group of Limnophilinae that share some characteristics, such as a flattened gonostylus with constriction at the distal end and a rough and chitinized anterior margin of the gonostylus.

Ribeiro (2008) found that the genus Eupilaria, which is used here as an outgroup (Fig. 7a), is not monophyletic, similar to several other genera in Limnophilinae. However, a part of Eupilaria, along with Idioptera, forms a clade with Eloeophila. Our preliminary analysis reinforces that both Euphylidorea and Phylidorea may also be paraphyletic (Alexander 1942). For example, our analysis suggests that Euphylidorea (Neophylidorea Petersen) bears many similarities to Phylidorea (Paraphylidorea Savchenko) (Fig. 2a) as noted by Alexander (1972) and Petersen et al. (2012). According to Starý (2009), Phylidorea, Idioptera and Eloeophila have similar aedeagal complexes defined as “second order parameres”, but Euphylidorea may also be included (Fig. 7b).

Clade 1. Synapomorphies: [2*: 1-0]; [13: 0-1]; [16: 0-1]; [22: 0-1]; [23: 0-1]; [25: 0-1]; [64: 1-2]. — This clade comprises the species of the Idioptera-Eloeophila complex. Although the close relationship between the two genera has long been recognized, our analysis now reveals that Idioptera is nested within Eloeophila. From now on, all species previously classified as Eloeophila will be referred to in the text as assigned to Idioptera.

This clade is distinguished mainly by a supernumerary m-cu [16: 0-1] (Fig. 3a). Edwards (1938) argued that the supernumerary cross-vein may be evanescent, disappearing in certain wings or even in both wings in some Idioptera specimens. Alexander (1948) noted that one specimen of I. angustior (Alexander, 1919) comb. nov. lacks the supernumerary m-cu in both wings. Species of unrelated genera such as the pediicid Nasiternella Wahlgren show a similar cross-vein, which is obviously of independent origin. Despite these exceptions, this feature is remarkable in the clade.

Apart from this supernumerary m-cu, this clade is supported by additional synapomorphies, including the lack of a spur at the base of vein Rs [2: 1-0]. However, this feature is highly homoplastic in Limnophilinae (Ribeiro 2008). Idioptera wings have a short discal cell [13: 0-1], although this character is reversed in Clade 11. Also, there is a new set of markings for this clade, one each around sc-r [22: 0-1] (Fig. 3a), around the humeral vein [24: 0-1] (Fig. 3a), and on distal margin of the arculus [25: 0-1] (Fig. 4a). We have found one synapomorphy related to the male terminalia: a curved contact between interbases [64: 1-2] (Fig. 8a). Ribeiro (2008) supported the monophyly of the clade Eloeophila + Idioptera based primarily on the lack of well-developed lateral processes of the aedeagal sheath.

Clade 2. Synapomorphies: [15: 0-1]; [52: 0-1]; [56: 0-1]; [58: 0-1]; [59*: 0-1]; [65*: 0-1]. — Three species of our sample are grouped within this clade: I. trimaculata (Zetterstedt, 1838) comb. nov., I. bifida (Alexander, 1921) comb. nov., and I. tergilobellus (Kato, 2021) comb. nov. Contrary to other species of Idioptera, the species of this clade have a gonostylus not laterally flattened [52: 0-1] (Fig. 9b-c). Another notable feature of this clade is the longer distance between the teeth of the clasper of the gonostylus [56: 0-1]. When the gonostylus is forked, the teeth are often close to each other. However, in Clade 2 they are distant from each other, even if the teeth remain short, such as in I. trimaculatacomb. nov.

Furthermore, members of this clade have the clasper with serrated margins, such as the anterior margin [59: 0-1] (Fig. 9b), which is homoplastic in Clade 13 and P. (Paraphylidorea). Another serrated section is the margin between the teeth of the clasper of the gonostylus [58: 0-1] (Fig. 9c). The distal bulge of the gonocoxite is more centralized in species of this clade, although it is unclear in I. bifidacomb. nov. This clade is also supported by another homoplastic feature: bent dorsal branch of the interbase [65: 0-1] which is also present in Clade 9 and Clade 13. The base of the clasper of the gonostylus also has a distinctive membrane with a different texture. However, it is difficult to observe this in I. bifidacomb. nov., and it may be a potential target for new morphological studies. The wings of the species in this clade have very similar markings and another synapomorphy: the distance between the tips of M₂ and M₃ is longer than the distance between the tips of M₃ and M₄ [15: 0-1].

Idioptera trimaculatacomb. nov. (Figs 3b; 9b) was frequently referred to as distinct from other species of the genus. Edwards (1938) highlighted the long antennae, different venation, and reduced spots (Fig. 3b). Despite sharing features with the original Idioptera (Clade 11), Edwards (1938) placed this species in the genus Eloeophila. Starý (2009) also pointed out that I. trimaculatacomb. nov. is an exceptional species of the genus, differing from the others. However, Ribeiro (2008) supported the monophyly of the genus Eloeophila, including I. trimaculatacomb. nov., but excluding other Idioptera species. Idioptera trimaculatacomb. nov. differs from the other two species of this clade mainly by the short teeth of the clasper and with a more serrated margin between them, and by having non-fused interbases.

Clade 3. Synapomorphies: [57: 0-1]; [63*: 1-0]. — This clade, which includes I. bifidacomb. nov. and I. tergilobelluscomb. nov., is sustained by a distinctive clasper of the gonostylus with a ventral tooth much longer than the dorsal tooth [57: 0-1] (Fig. 9c). A significant difference between the two species of the clade is the short lobe of the gonostylus of I. tergilobelluscomb. nov. Although the wing of these species appear to be unmarked, there remains some markings (Fig. 3a). Idioptera tergilobelluscomb. nov. has interbases fused [63: 1-0]. Idioptera bifidacomb. nov. appears to have similar interbases. Both species have a dilated distal bulge on the gonocoxite, as well as a nearly subterminal gonostylus. The most noticeable feature shared between these species is the presence of dorsal extensions (paired lobes) on tergites, as evidenced in the description of I. bifidacomb. nov. and in Kato (2021). The ninth tergite has a bifid central projection, with the more distal lobe being more evident in I. bifidacomb. nov.

Clade 4. Synapomorphies: [0*: 1-0]; [18: 0-1]; [26: 0-1]; [29: 0-1]; [54: 0-1]. — This clade is supported by a set of seven markings on the wing. In an analysis of Palearctic Eloeophila species, Starý (2009) already highlighted the presence of “seven larger spots on the anterior margin of the wing”. This heptad pattern proposed by Edwards (1938) is composed of the following markings: around the humeral vein; between humeral vein and origin of Rs; on the base of Rs; around sc-r; pterostigma; on tip of R₃; on tip of R₄. Of the seven markings, six were already present. Only the marking between the humeral vein and origin of Rs [26: 0-1] (Fig. 4a) is a synapomorphy for this clade. These seven markings on the anterior margin of the wing are darker and larger, making them the most prominent. However, they may be weaker in certain extinct Idioptera such as I. moba (Podenas) comb. nov. and I. eocenica (Santos et al.) comb. nov.

The clade is also supported by the ventral margin of the clasper of the gonostylus pointed medially [54: 0-1] (Fig. 9j), and by two wing features: an arched tip of the anal vein [18: 0-1] and a marking at the tip of R₄ [29: 0-1] (Fig. 4a). This clade also has short antennae [0: 1-0], a reversal that also has changed in Clade 11. The size of the antennae is an important feature in identification keys and diagnoses. However, males have longer antennae compared to females in this genus. In species designated as having “long antenna”, as long as the length of the head and thorax combined, this feature is more evident in the males. The antennomeres of Idioptera have an almost rounded shape, but in some males, like in Clade 11, the oval flagellomeres are longer.

Sexual dimorphism is also present in the wing markings. Male wings may have stronger markings than female wings. This characteristic may lead to errors such as false new species based on “markless” wings. There is also variation in the color and intensity of the markings (Starý 2009), which requires a careful examination of designated synonyms. Many regional variations have been designated as subspecies or forms, and numerous species have been synonymized without thorough revision. Although not every variation is a distinct species, some defined species may consist of different lineages blended and amalgamated. While a comprehensive revision of the material is beyond the scope of this study, we acknowledge the necessity for such detailed examinations.

Clade 5. Synapomorphies: [3: 0-1]; [12: 0-1]; [14*: 1-0]. — This clade includes the extinct species of Idioptera (all from the Eocene). One species (I. scudderi (Santos et al.) comb. nov.) is from the Florissant Formation in Colorado, USA (Scudder 1894), and two species (I. moba and I. trimaculata) are from Baltic Amber. The clade is supported by the downward inclination of vein Rs [3: 0-1] (Fig. 4). The clade has two wing synapomorphies: a non-inclined tip of the vein M₄ [12: 0-1] and m-m shorter than half of the basal deflection of M₃ [14: 1-0]. The straight M₄ is a feature shared with an outgroup species, Phylidorea (Macrolabina) nigronotata (Siebke, 1870). The length of m-m is a more homoplastic character, with reversions outside the ingroup, such as in the outgroup species Phylidorea (Phylidorea) pietatis (Alexander, 1950), Clade 5, Clade 9, and I. sparsipunctum (Starý, 2009) comb. nov.Podenas (2003) highlighted that I. mobacomb. nov. (Fig. 4b) is smaller than other Idioptera, measuring 6–7 mm. This species is similar in size to I. eocenicacomb. nov. and I. scudderi (Santos et al.) comb. nov. (Fig. 4a). Podenas (2003) also stated that I. mobacomb. nov. (Fig. 9a) shares more terminalia features with I. trimaculatacomb. nov. than with other Idioptera. However, these resemblances may be due to symplesiomorphies.

Finding synapomorphies related to the morphology of the male terminalia is complicated: only females of I. eocenicacomb. nov. were found; the male terminalia of I. scuddericomb. nov. is unclear; and the aedegal complex of I. mobacomb. nov. is not completely shown. These three species were revised by Santos et al. (2022).

Clade 6. Synapomorphy: [5: 0-1]. — This clade groups I. eocenicacomb. nov. with I. mobacomb. nov., both from Baltic Amber. Species in this clade are singletons, meaning they are only known from a single specimen each. The wings of these species have a tiny vein R₂ [5: 0-1]. The male terminalia of I. mobacomb. nov. has an apomorphic ninth tergite with a convex middle tergite instead of the typical concavity. Starý (2009), however, stated that the shape of the ninth tergite is not precise and reliable for identifying subgroups in this genus. Unfortunately, there are no visible aedeagal structures such as parameres or interbases. Another problem in understanding the relationship between these species is that in I. eocenicacomb. nov., only the female was found, while for I. mobacomb. nov., only the male is known.

Podenas (2003) stated that wing of I. mobacomb. nov. (Fig. 4b) resembles I. trimaculatacomb. nov. but it is more due to plesiomorphies shared between them. Santos et al. (2022) highlighted the similarities among these Baltic Amber fossil species. Both species of Clade 6 have faint markings. Podenas (2003) only described the pterostigma for I. mobacomb. nov., although a thorough analysis may reveal the homologous markings (Fig. 4b). Other species of Idioptera have already been described as markless, such as I. johnsoni (Alexander, 1914) comb. nov. This highlights the importance of conducting a comprehensive analysis of the specimens to identify the subtle yet significant structures.

Clade 7. Synapomorphies: [2*: 0-1]; [21: 0-1]; [33: 0-1]; [34: 0-1]; [38: 0-1]; [60: 0-1]; [62*: 0-1]. — This clade is supported by a series of markings on the wing (Fig. 5a b). There is a marking on the distal portion of the discal cell [34: 0-1] (Fig. 5a) and a marking at the base of the wing [21: 0-1] (Fig. 5a). The marking at the base of the wing is most noticeable in certain Idioptera (Clade 11) because of the dark color of the wing markings in this clade. Furthermore, there are some small markings near the tips of the medial and cubital veins [33: 0-1]. These distal rounded markings are more easily distinguished in certain species, but they are consistently present in the clade. Sometimes the markings on the tips of M₁ and M₂ are so large that they may fuse with the markings on the tips of R₄ and R₅. Another apomorphy in the wing is the presence of a double marking on the anal vein [38: 0-1] (Fig. 5a). The clade is also supported by the presence of a spur at the base of Rs [2: 0-1] (Fig. 5b) with a reversion in Clade 15.

The clade is also defined by features in the male terminalia. The most remarkable feature is the presence of a crest on the anterior margin of the clasper of the gonostylus [60: 0-1] (Fig. 9h). This crest may be highly developed (see Clade 9) or less visible due to a dorsal bend (Starý, 2009). The clasper of the gonostylus in this group is described as rectangular-shaped, flattened and blackened (Edwards 1938). Podenas et al. (2020) defined this as a “blade-shaped” clasper of gonostylus. The flattened shape is a plesiomorphy, but species of Clade 7 have a distinctively darkened rectangular clasper. The dorsal tooth of the clasper of gonostylus is bent downwards [62: 0-1] (Fig. 9d–j), which reinforces the rectangular shape of the clasper. At least, this clade lacks the distal bulge in the gonocoxite [50: 1-0], present in the outgroups.

A feature common in this clade is a great concentration of setulae on the tip of the wing. Alexander (1938) pointed out that species, such as E. seticellula (Alexander, 1938), may be different from other Eloeophila. This species lacks patches and a marking on R₅ but has a great concentration of wing setulae. He created a genus, Trichephelia Alexander, to cover this setose species. The presence of these setulae is sometimes hard to observe and delimit. Alexander (1957) stated that this feature is weak for diagnosis and posteriorly synonymized Trichephelia with Eloeophila. The presence of species with many setae in the wing was reinforced, for example, in the description and naming of E. apicisetula (Kato, 2021). Further analysis is needed to understand if the trichiation is a stable feature and if the group of species around I. pusillacomb. nov. in Clade 7 is equivalent to Alexander’s Trichephelia genus. For now, we reinforce the synonym of Trichephelia under Eloeophila.

Clade 8. Synapomorphies: [20: 0-1]; [37: 0-1]; [53: 0-1]; [65*: 0-1]. — This clade is supported by two wing features: the thinning of vein C between veins R₄ and R₅ [20: 0-1], and a marking at the base of vein CuA [37: 0-1 (Fig. 5b)]. There are also two male terminalia features: a dorsal spur at the base of the clasper of the gonostylus [53: 0-1] (Fig. 9e), and the distal branches of the interbases parallel to the aedeagus [65: 0-1] (Fig. 8c). The straight branch of the interbase is also shared with Eupilaria phoenosoma (Fig. 7a) (outgroup). Despite being enlarged in some species, such as I. nearctica (Alexander, 1966), the dorsal branch retains a straight side, which may be notable for species in Clade 8.

Clade 9. Synapomorphies: [14*: 1-0]; [23: 0-1]; [54: 0-1]; [61: 0-1]; [65*: 0-1]. — This clade is defined by a pointed spur in the crest of the clasper of the gonostylus [61: 0-1] (Fig. 9e). In the species of this clade, the crest is tall and extends into a spur. The size of this spur is variable, and it may only be visible by an elevation in the crest, as seen in I. aldrichi (Alexander, 1927) comb. nov., or even as tall as the remainder of the clasper such as in I. kintaro (Alexander, 1957) comb. nov. (Fig. 9e). Also, the clade is supported by a bent tip of the dorsal branch of the interbase [65: 0-1], clasper non-bifurcated [54: 0-1], and a short m-m [14: 1-0]. An additional synapomorphy for this group is a marking at the tip of Sc, separate from the large marking around sc-r [23: 0-1] (Fig. 5b). Species in this clade also have a narrow wing stalk.

Clade 10. Synapomorphies: [32: 0-1]; [41: 0-1]. — This clade is supported by two wing markings: a marking near the tip of R₅ [32: 0-1] and a patch that connects the pterostigma with the marking around the basal deflection of R₅ [41: 0-1] (Fig. 6a-b). These connecting patches are not exclusive to this clade. Other craneflies, such as Dicranophragma (Dicranophragma) perlatum (Alexander, 1955), may exhibit a similar pattern. However, it is an apomorphy in the phylogenetic context of the genus.

Clade 11. Synapomorphies: [0*: 0-1]; [6*: 0-1]; [10*: 0-1]; [13: 1-0]; [17: 0-1]; [39*: 0-1]; [43*: 0-1]; [46: 0-1]; [51: 0-1]. — This clade includes the species that were originally included in Idioptera. The inclusion of Idioptera within Eloeophila makes the latter paraphyletic. The definition of the genus Idioptera has always been controversial. Macquart (1834) established this genus based on the simultaneous presence of long antennae and supernumerary m-cu. Further diagnosis for this genus includes features present in other groups: supernumerary m-cu (as in Eloeophila), gonostylus without a serrated margin (as in Phylidorea), tip of R₁ up-curved (as in Euphylidorea), antenna slender (as in I. trimaculatacomb. nov.), long discal cell (as in Phylidorea), and banded wings (as in I. apicatacomb. nov.).

Edwards (1921) merged Idioptera with Eloeophila, but later (Edwards 1938) he redefined them as separated entities. The lack of a phylogenetic background for this group has obscured the true synapomorphies of this clade. Edwards (1938) argued that certain Idioptera species resemble Phylidorea in some aspects of the terminalia, the long discal cell and the long antenna. However, the supernumerary m-cu and wing markings are evidence of a concise Idioptera. The similarities between certain Phyllidorea species with the Eloeophila + Idioptera complex are also discussed by Alexander (1921).

Despite this, Clade 11 is strongly supported in our analysis, reinforcing the monophyly of the group that was originally named Idioptera. A recurring diagnostic feature is the banded wings (Fig. 6a). This banded appearance is caused by the presence of large patches between the large markings. This group has an additional patch: one connecting the marking around m-cu with the marking at the tip of vein CuA [43: 0-1]. The banded appearance is reinforced by a distinctive feature. The wing patches of members of Clade 11 are enlarged and have a light core with a dark frame [46: 0-1]. The markings in this species of this clade are frequently larger than those in other groups, which supports the idea of a banded shape. These large transverse bicolor patches are very diagnostic for this group.

Two characters frequently observed as diagnostic features of this group are a long discal cell [13: 1-0] and a long antenna [0: 0-1]. They are considered as synapomorphies of this clade, although they may exhibit plasticity in Idioptera. Furthermore, this group is characterized in the wing by a supernumerary m-cu inserted proximal to the level of the origin of Rs [17: 0-1], the two anal vein markings far apart [39: 0-1], short R₂₊₃₊₄ [6: 0-1], and distal M₁₊₂ is longer than twice the length of M₂ [10: 0-1]. The male terminalia are distinguished from the other groups by the presence of a curved gonostylus [51: 0-1] (Fig. 9h). Despite its curved shape, the clasper of the gonostylus retains a flattened aspect.

Clade 12. Synapomorphy: [44: 0-1]. — This clade represents the Palearctic species of the original genus Idioptera. The clade is defined by the presence of a patch connecting the marking around m-cu with the marking at the tip of vein CuP [44: 0-1] (Fig. 6a). This additional patch reinforced the banded appearance of Idioptera wings. These “banded wings” (Fig. 6a) were identified as a diagnostic feature of the original genus Idioptera.

Clade 13. Synapomorphies: [59*: 0-1]; [64: 0-1]; [71: 0-1]. — This clade contains the bulk of the diversity of the entire Idioptera-Eloeophila complex. The group is supported by the serrated margin of the bulge in the base of the gonocoxite [59: 0-1] (Fig. 8d). The ventral tooth of the clasper of the gonostylus is more rounded in this clade. The clade also has a bent tip of the dorsal branch of the interbase [64: 0-1] and humps on the lower shelf [71: 0-1]. These membranous projected humps were evidenced by Starý (2009). He referred to them as a lateral projection of the lower shelf (“lpsh”). To avoid confusion with the lateral projection of the sheath of the aedeagus, we will refer to it as the humps of the lower shelf.

Clade 14. Synapomorphies: [40*: 0-1]; [42: 0-1]; [45: 0-1]; [72: 0-1]. — This clade groups the banded wing I. apicatacomb. nov. with the spotted wing Eloeophila. This clade is supported by a patch connecting the marking around sc-r with the marking around the basal deflection of R₅ [42: 0-1] (Fig. 6b). This results in a characteristic Y-shaped pattern in the distal part of the wing. Another patch of this group is located between the marking around the humeral vein and the marking at the base of CuA vein [45: 0-1]. Although the original genus Idioptera (Clade 11) was referenced as having banded wings, the species in Clade 14 have more patches and “bands” on their wings.

Additionally, this clade is supported by a marking on the anal lobe near the margin [40: 0-1] (Fig. 6b). The clade is also defined by the presence of a median projection of the lower shelf [72: 0-1] (Fig. 8b). As defined by Starý (2009), this pointed projection (“mplsh”) emerged from the lower shelf in the direction of the aedeagus. It may be an extension of the anterior margin of the vesica. In some species, such as I. maculatacomb. nov., the mplsh are shorter and more rounded.

Clade 15. Synapomorphies: [2*: 1-0]; [36: 0-1]; [47: 0-1]. — This clade groups the Idioptera with spotted wings. The main feature of this group is the presence of spots. The spots are a set of numerous small dark rounded markings [47: 0-1] (Fig. 6b). Mainly, the spots are located along the longitudinal veins, but they may also be found on the anal lobe and in the costal cell, such as in I. granulata (Edwards, 1926) comb. nov.Schiner (1863) utilized the German term “buntgefleckt” (translated as “spot” or “freckle”) to describe these small rounded marks. He described the genus Ephelia, to group species of spotted wings. The term “spots” is also employed in subsequent studies such as Starý (2009).

Schiner’s (1863)Ephelia was synonymized under Eloeophila and is now synonymized under Idioptera. Besides the spots, Clade 15 species have a long marking around m-cu [36: 0-1] and lack a spur at the origin of Rs [2: 1-0] (Fig. 6b). Edwards (1938) redefined Eloeophila as a combination of a supernumerary m-cu vein and wings with spots while still classifying it as a subgenus of Limnophila Macquart. This definition was adopted by Dienske (1987) and other publications. This traditional genus definition is closely linked to the group of species with spotted wings. The number and intensity of wing spots may vary.

Clade 16. Synapomorphy: [19: 0-1]. — Members of this clade share a spur at the tip of the anal vein [19: 0-1] (Fig. 6b). Some species, like I. linnei (Oosterbroek, 1992), have a spur distally on the anal vein. However, the spur of the wings of this clade is characterized by a distinctive orientation, directed towards the anal lobe. This spur is present in the base of the arched tip of the anal vein but does not reach the posterior margin of the wing and it is more evident in males.

Clade 17. Synapomorphy: [1: 0-1]. — Members of this clade have large wings. This enlarged shape (Fig. 6b) is due to an extended posterior margin in the middle of the wing [1: 0-1], as seen in I. maculatacomb. nov. (Fig. 6b). This feature is more evident in the males.

A phylogenetic analysis of the Idioptera-Eloeophila complex was conducted using a data matrix consisting of 25 terminal taxa and 73 characters. The selection of terminal taxa was not arbitrary, but rather based on a meticulous review of available morphological information in the taxonomic literature, as well as direct observation of specimens from both the ingroup and outgroup. The resulting data matrix effectively captures the full range of morphological variation observed within the study group. Based on our findings, we reexamined the diversity of morphologies and we placed each species in the evolutionary background.

The paraphyly of Eloeophila generates a taxonomic problem. There are two solutions for this scenario. A solution for this would be to elevate six clades to genus level: Clade 2 (species related to I. trimaculatacomb. nov.), Clade 5 (fossil species related to I. mobacomb. nov.), the old genus Trichephelia (species related to I. pusillacomb. nov.), Clade 9 (species related to I. aldrichicomb. nov.), Clade 11 (original concept of Idioptera), and Clade 13 (old genus Ephelia). The other solution is to synonymize both valid genera under the same genus. We chose the last option to retain the conspicuous synapomorphy for this genus: the distinctive supernumerary m-cu and the better supported analysis. Further studies with better internal representation may support or not the independence of the six lineages included in Idioptera.

Although Eloeophila is much more diverse, the oldest name is Idioptera. Eloeophila was described by Rondani in 1856, while Idioptera was described by Macquart in 1834. From a taxonomic perspective, the older name Idioptera must be used to comply with the Principle of Priority of the Code. So, the entire lineage must be Idioptera. This act requires 89 taxonomic changes, which we have summarized below.

Based on this analysis and the acceptance of Eloeophila as a junior synonym of Idioptera, we can establish a phylogenetic framework that predicts the most probable classification of all described species (over 80 species) into the clades illustrated in Figure 1. Such a scheme is informed by the distribution of the derived character states, as revealed by a thorough examination of the descriptions of all the species and reinforced the stability of main clades present in the now expanded concept of the genus Idioptera. However, it is important to note that this framework should be considered provisional and subject to further validation through additional testing.