To ensure comprehensive representation of all Gripopterygidae subfamilies, our study included multiple genera and species. Specimens from South America (including the Neotropical region, the South American transition zone, and the Andean region (sensu Morrone (2014, 2015)), as well as from Australia, were examined at the Centro de Investigación Esquel de Montaña y Estepa Patagónica (CIEMEP) in Esquel, Chubut, Argentina, and at the Collection of Aquatic Insects “Prof. Dr. Cláudio Gilberto Froehlich”, Aquatic Biology Laboratory, State University of São Paulo (UNESP), Assis, São Paulo, Brazil (Table 2). Additional Australian specimens were examined at the Stuttgart State Museum of Natural History (SMNS), Stuttgart, Germany.

Morphological characters and primary homologies were supplemented using key original works, including Illies (1963, 1965), Froehlich (1969, 1990, 1994, 1998, 2001), McLellan (1977), and McLellan and Zwick (2007). Wing venation terminology follows Béthoux (2005).

In this work we included specimens from the following institutions: CIACGF – Collection of Aquatic Insects “Prof. Dr. Cláudio Gilberto Froehlich”, Aquatic Biology Laboratory, State University of São Paulo (UNESP), Assis, São Paulo, Brazil; CIEMEP – Centro de Investigación Esquel de Montaña y Estepa Patagónica, Esquel, Chubut, Argentina; MZUSP – Museum of Zoology of the University of São Paulo, São Paulo, Brazil; SMNS – Stuttgart State Museum of Natural History, Stuttgart, Germany.

The specimens’ external morphology was examined by the authors using a Leica S9E and a Leica M205A stereomicroscope. The photographs were taken using a Leica M205A stereomicroscope and processed using the image editing software Adobe® Photoshop CC. The line drawings were prepared with the aid of a camera lucida and vectored on Adobe® Illustrator CC, as well as the final topology.

We used 41 terminal taxa, with 38 (29 genera) falling under Gripopterygidae and three (two genera) falling under Austroperlidae (Table 2), in line with the recent studies that recovered a sister relationship between these families (see McCulloch et al. 2016; Ding et al. 2019; Letsch et al. 2021; García-Girón et al. 2024). The dataset was built using Mesquite 3.81 (Maddison and Maddison 2023) and exported to the format employed in the ‘Tree analysis using New Technology’ software (TNT v. 1.6; Goloboff and Morales 2023). The dataset is provided in Table S1.

Penturoperla barbata Illies, 1960 was selected as the root of the tree. For the 41 terminal taxa, we selected 50 morphological characters (28 binary and 22 multistate) and their associated states (see section “Morphological Characters List”). Character coding followed the contingency method proposed by Sereno (2007), with four logical components (locator, variable, variable qualifier, and character status), and the criteria of character types as proposed by Simões et al. (2016). In the dataset, we used a “?” when a character could not be found in each species and a “–” when a character was deemed inapplicable.

The Parsimony analysis was utilized, treating all characters as nonadditive (Fitch 1971). We employed the “Traditional search” command in TNT, with Tree Bisection Reconnection (TBR) branch swapping, conducting 1,000 replications, saving 100 trees per replication, and collapsing trees after the search.

We performed an initial analysis using Equal Weighting (EW) Parsimony. However, EW usually produces a consensus tree that is overly conservative, failing to accurately represent phylogenetic relationships (Goloboff et al. 2008, 2018). To address this limitation, we focused on identifying characters with greater cladistic congruence while downweighing those with a higher degree of homoplasy (Goloboff 1997). Analyses were subsequently conducted under Implied Weighting (IW) Parsimony to account for homoplastic characters (Goloboff 1993; Giribet 2003; Goloboff et al. 2008). IW employs a parameter (k, the concavity constant) to determine the degree to which homoplasy (independently evolved similar traits) is penalized during phylogenetic reconstruction. Goloboff (1993) originally introduced the method with a strong concavity (k = 3), which remains the default in the TNT software (Goloboff et al. 2018). However, later studies, particularly those involving larger datasets, showed that better results have been obtained with higher k values (e.g., k = 12; see Goloboff et al. 2008, p. 765), which apply milder penalties to homoplastic characters (Goloboff et al. 2008, 2018; Smith 2019). Following these recommendations, we employed a range of k values, from 3 to 15, to evaluate potential topological changes and assess the impact of weighting against homoplasy. We also analyzed the dataset in TNT with the option “choose k value so that the weight ratio between no homoplasy and maximum possible steps is the ratio 1 to 10,000” to identify a potential optimal k value for the dataset under analysis.

To estimate branch support, we employed Relative Bremer support (Goloboff and Farris 2001), utilizing 1,000 suboptimal trees up to five steps longer (obtained through the traditional search). This approach provided a robust measure of support for various branches in the resulting cladogram.

Characters and their states (e.g., char. 25, state 1: [25:1]) or the transformation sequence (e.g., char. 25, state 0 to 1: [25:0–1]) are mentioned in the text if needed.

Male and female specimens that were collected during mating in the field are hereby considered probable new species. Their placement within the subfamily Gripopteryginae is examined and tested. The new name presented herein conforms to the International Code of Zoological Nomenclature and has been registered in ZooBank, the ICZN’s online registration system.

We studied the morphology of Gripopterygidae genera, focusing on the subfamily Gripopteryginae. As a result, we selected and encoded 22 nymphal morphological characters (characters 1–22) and 28 adult morphological characters (characters 23–50), reaching 50 characters. The list presented below includes characters proposed by Illies (1963) and McLellan (1977) for the subfamilies of Gripopterygidae.

Nymphs – Body and head (characters 1–4)

1 Body; setae: (0) small spine-like, (1) vesicular-like (Fig. 2A), (2) claviform-like (Fig. 2B), (3) hook-like (Fig. 2C). (In Froehlich 1969, 2001).

2 Body; debris covering surface: (0) absence, (1) presence. (In Illies 1963).

3 Head; antennae, ring of curved hair-like setae at least on antennae basal third: (0) absence, (1) presence.

4 Head; antennae, fringe of hair-like setae: (0) absence, (1) presence (Fig. 2D).

Nymphs – Thorax (characters 5–12)

5 Pronotum; surface, cuticle structure: (0) smooth (without small processes), (1) with small spine-like processes or elevations (Fig. 2E). (In Illies 1963).

6 Pronotum; anterior corners, shape: (0) corners rounded, (1) corners angulated, (2) corners with strong spines, (3) corners with strong projection (Fig. 2F and Fig. 3A).

7 Pronotum; lateral margin, extension (paranota): (0) absence, (1) presence (Fig. 2F). (In Froehlich 2001).

8 Metanotum; mid-distal margin (between wing pads), shape: (0) circular, concave (Fig. 2G), (1) circular, convex (Fig. 2H), (2) angular excised (pyramid-shaped) (Fig. 2I), (3) angular (V-shaped) (Fig. 2J), (4) straight, (5) bilobated (W-shaped) (Fig. 2K).

9 Coxa; procoxal process: (0) absence, (1) presence (Fig. 2D). (In McLellan et al. 2005).

10 Femur; extensor margin, fringe of hair-like setae, shape: (0) absence, (1) very dense fringe of hair-like setae (Fig. 2D), (2) sparse fringe of hair-like setae. (In Illies 1963; McLellan 1977).

11 Femur; extensor margin, small spine-like setae: (0) absence, (1) presence (Fig. 2L).

12 Tarsus; ventral surface of the third tarsal segment, type of setae: (0) setiform, small, thin setae (Fig. 2M), (1) thick, hair-like setae (Fig. 2N). (In McLellan and Zwick 2007).

Nymphs – Abdomen (characters 13–22)

13 Abdomen, segments 4 to 9, constriction: (0) unconstrained, sides approximately parallel, (1) constrained on its middle section, (2) unconstrained, sides not parallel, segment thicker on its apical section.

14 Abdomen; terga 1–9, row of mid-dorsal hair-like setae: (0) absence, (1) presence.

15 Abdomen; terga 1–9, row of mid-dorsal processes/spines: (0) absence, (1) one row (Fig. 2O), (2) two rows (Fig. 3B, C). (In Illies 1963).

16 Abdomen; distal margin of terga 1–10, type of setae: (0) setiform, small, thin setae, (1) setiform, stout setae, (2) spine-like setae, (3) claviform-like setae, (4) vesicular-like setae, (5) long, hair-like setae.

17 Abdomen; tergum 10, shape of distal margin: (0) curved, (1) trapezoidal, (2) triangular, (3) acuminated.

18 Abdomen; anal gills: (0) absence, (1) presence (Fig. 2O).

19 Abdomen; modified terminal structures (paraprocts) for breathing: (0) absence, (1) presence. (In McLellan 2001b).

20 Paraprocts; distal region (apex): (0) rounded (Fig. 2P), (1) with a long spine (Fig. 3C), (2) acute. (In Vera 2006a; Vera 2009).

21 Cerci; cercomeres, ring of small distal setae: (0) absence or weakly developed, (1) presence.

22 Cerci; long hair-like setae, shape: (0) absence, (1) scattered arrangement, (2) fringed arrangement.

Adults male and female – Head (characters 23–24)

23 Ocelli: (0) absence, (1) presence.

24 Head; maxillary palp, size of the fifth segment relative to the third: (0) subequal (nearly equal), (1) two times longer or more. (In Vera 2009).

Adults male and female – Thorax (characters 25–37)

25 Wing; state: (0) apterous, (1) micropterous or brachypterous recorded in male or female, (2) macropterous (Fig. 5C). In some widespread Patagonian species, micropterism or brachyperism is common and variable through the distributional range; we codify it as state 1 if the character state has been recorded in at least a few specimens. Some widespread species are always macropterous (Pessacq & Duarte pers. obs., but also see Illies, 1963).

26 Wing; fore and hind wing pigmentation patterns: (0) both wings unpigmented (Fig. 5C), (1) forewing pigmentation, hind wing unpigmented, (2) both wings pigmented. — Character not applicable to taxa with state 0 in character 25.

27 Forewing; crossveins between C and Sc, number: (0) one crossvein (Fig. 5C), (1) more than one crossvein (Fig. 3D). — Character not applicable to taxa with state 0 in character 25.

28 Forewing; pterostigmatic cell, crossveins: (0) absence (Fig. 5C), (1) presence. (In Illies 1963; Froehlich 1969; McLellan 1977). — Character not applicable to taxa with state 0 in character 25.

29 Forewing; RP vein, shape: (0) unforked, (1) one fork (Fig. 5C), (2) two forks. (In Illies 1963; McLellan 1977). — Character not applicable to taxa with state 0 in character 25.

30 Forewing; anterior fork of RP vein relative to RA: (0) anterior fork of RP vein not connected to RA (Fig. 5C), (1) anterior fork of RP vein connected to RA. (In Froehlich 1969; Vera 2016). — Character not applicable to taxa with state 0 in character 29.

31 Forewing; CuA vein, shape: (0) unforked, (1) one fork (Fig. 5C), (2) two forks. (In Froehlich 1969; McLellan 1977). — Character not applicable to taxa with state 0 in character 25.

32 Hind wing; inferior branch of the M vein relative to CuA vein, shape: (0) unfused, (1) partially fused, separating near the wing margin (Fig. 5C), (2) completely fused. (In McLellan 1977).

33 Hind wing; sixth anal vein, shape: (0) sixth anal vein free of wing margin, (1) sixth anal vein fused to wing margin (Fig. 5C). (In McLellan 1977).

34 Femur; flexor margin, distal spine: (0) absence, (1) presence (Fig. 2Q, R). (In Froehlich 1969; Froehlich 2001).

35 Tibia; distoventral region, spurs: (0) absence, (1) presence (Fig. 2R). (In Froehlich 1969; McLellan 1977).

36 Tarsus; basal tarsal segment relative to apical segment: (0) basal segment as one-third of the length of the apical segment, (1) basal segment as half the length of the apical segment (Fig. 2R), (2) basal segment about as long as the apical segment. (In Illies 1963; Vera 2016).

37 Tarsus; ventral side of the basal segment, narrow membranous band: (0) absence, (1) presence. (In Vera 2009).

Adult male – Abdomen (characters 38–50)

38 Abdomen; tergum 10, anterior sclerites with oval tubercles: (0) absence, (1) presence.

39 Abdomen; tergum 10, anterior sclerites, fusion (inner margin): (0) completely or partially (at least its anterior half) fused, (1) touching at a small point or separated by a narrow membrane. (In Pessacq et al. 2020).

40 Abdomen; tergum 10, anterior sclerites relative to central sclerite: (0) completely fused with central sclerite (Fig. 2S and Fig. 5D), (1) separated from central sclerite by a suture, (2) partially fused with central sclerite, with a narrow lateral cleft (considering fusion of central and lateral sclerites), (3) separated from the central sclerite by a membrane (Fig. 2T).

41 Abdomen; tergum 10, central sclerite, distal margin, shape: (0) not protruding, (1) protruding in a simple lobe, (2) protruding with lobe divided in two parts (Fig. 5D).

42 Abdomen; tergum 10, central sclerite, denticles/teeth at the distal margin: (0) absence, (1) one denticle, (2) two denticles (Fig. 5D–F).

43 Abdomen; tergum 10, posterior sclerite clearly differentiated, base surrounded by a membrane: (0) absence (posterior sclerite fused to central sclerite), (1) presence (Fig. 2U). (In Illies 1963; McLellan 1977).

44 Abdomen; posterior sclerite, shape: (0) small spine/knob, (1) elongated, digitiform, (2) large, rounded, (3) small, rounded. — Character not applicable to taxa with state 0 in character 43.

45 Abdomen; posterior sclerite, position: (0) apical, (1) ventral. — Character not applicable to taxa with state 0 in character 43.

46 Abdomen; epiproct with a ventral sclerotized projection: (0) absence (Fig. 2V and Fig. 5E, F), (1) presence (Fig. 2U).

47 Abdomen; epiproct, shape: (0) concave, spoon-shaped, (1) flat, tip projected ventrally, (2) digitiform, curved, (3) digitiform, straight, with a wide base. (In Illies 1963). — Character not applicable to taxa with state 0 in character 46.

48 Abdomen; epiproct, lobe in the ventro-basal region: (0) absence, (1) presence. (In Illies 1963). — Character not applicable to taxa with state 0 in character 47.

49 Abdomen; epiproct, keel in the ventro-basal region: (0) absence, (1) presence. — Character not applicable to taxa with state 0 in character 46.

50 Abdomen; epiproct, inner surface, number of denticle rows: (0) absence, (1) one row of denticles, (2) two rows of denticles. — Character not applicable to taxa with state 0 in character 46.

The analyses conducted produced different consensus topologies based on the methods applied: (1) equal weighting (EW) (Fig. S1) and (2) implied weighting (IW) with varying k values (k = 3, 5, 7, 9, 11, 13, 15) (Fig. 4 and Figs S2–S3). As expected, and mentioned in section 2.4, the EW analysis resulted in a fully polytomous consensus topology based on 10 most parsimonious trees (MPTs) and will be no further discussed.

In contrast, the IW analyses consistently outperformed EW, recovering four well-resolved consensus topologies across the evaluated k values: k = 3, with a consensus of 66 trees; k = 5, with 27 trees; k = 7, with 9 trees; and k = 9–15, with 27 trees. For the dataset being studied in TNT, k = 9.60936 was the possible optimal k value. For this value, the consensus topology was like that of k = 9–15.

The general structure of the consensus topologies and the synapomorphies regarding Gripopteryginae taxa stayed constant, despite slight differences in the relationships between some of the taxa. K values between 9 and 15 resulted in identical topologies and synapomorphies (Fig. 4), a k value of 7 resulted in a slightly different tree (i.e. Neopentura semifusca appears in a clade together with Pehuenioperla llaima instead of a politomy, see Fig. 4 and Fig. S3), while k values of 3 and 5 resulted in completely different topologies, but with a monophyletic Gripopteryginae. From this point onward, the results will be based on the k values ranging from 7 to 15, reinforcing this decision with the optimal k value obtained with TNT (k = 9.60936). For the same reason, discussion will be based on the results obtained with k = 9.60936. However, the trees resulting from analyses with the remaining k values can be consulted in Figs S2–S3.

The morphological phylogenetic analyses of Gripopterygidae consistently yielded the family as monophyletic across all weighting schemes (for Relative Bremer support, see Fig. 4). Five synapomorphies supported this classification (e.g., [18:1], nymphal stage with abdominal anal gills; [19:0], nymph without terminal structures (paraprocts) modified for breathing; [32:1], adult hind wing with inferior branch of the M vein partially fused to the CuA vein and separating near the wing margin; [36:2], adult basal tarsal segment about as long as the apical segment; and [49:0], absence of keel in the ventro-basal region of epiproct) (Fig. 4).

The subfamily Gripopteryginae (Clade A, Fig. 4) was recovered as polyphyletic due to the inclusion of Neopentura semifusca Illies, 1965 (a genus whose position has been uncertain since its description, as will be discussed later) outside the subfamily’s delimitation. The remaining Gripopteryginae taxa consistently formed a clade (Clade A, Fig. 4), supported by distinct synapomorphies. For k = 9.60936 (and similarly for k = 9–15), the following characters states supported the clade: the forewing CuA vein with one fork [31:1]; the sixth anal vein of the hind wing fused to the wing margin [33:1]; a simple, protruding lobe on the distal margin of the central sclerite in tergum 10 [41:1]; and the absence of a posterior sclerite in tergum 10 [43:0] (Fig. 4).

The inclusion of Neopentura semifusca (a species currently classified in Gripopteryginae) along with the clade of Antarctoperlinae resulted in the subfamily being consistently paraphyletic. For k = 9.60936, Neopentura semifusca (Fig. 4) appeared in a polytomy with Pehuenioperla llaima Vera, 2009 (a typical Antarctoperlinae) and the remaining Antarctoperlinae taxa, while under k = 7, it appeared as the sister to Pehuenioperla llaima (Fig. S3). The synapomorphies that supported the clade are: the absence of a fringe of hair-like setae on the nymphal extensor margin of the femur [10:0]; the presence of thick, hair-like setae on the ventral surface of the third tarsal segment [12:1]; the presence of a long spine on the distal region (apex) of the nymphal paraprocts [20:1]; the absence or weak development of a ring of small distal setae on the nymphal cercomeres [21:0]; the absence of spurs in the tibia distoventral region [35:0]; and a simple, protruding lobe on the distal margin of the central sclerite in tergum 10 [41:1] (Fig. 4).

Dinotoperlinae and Leptoperlinae (sensu McLellan 1977) were consistently recovered as polyphyletic. Regarding Dinotoperlinae, Trinotoperla irrorata Tillyard, 1924, appeared at the base of Gripopterygidae, while Dinotoperla serricauda Kimmins, 1951, and Alfonsoperla flinti McLellan & Zwick, 2007, were recovered in a polytomy (Figs 4, S3).

Leptoperlinae also appeared as polyphyletic, divided in three different and separated clades. Newmanoperla thoreyi (Banks, 1920) was consistently grouped with Notoperla species, supported by [4:1], the presence of a fringe of hair-like setae on the nymphal antennae; [16:1], the presence of a row of mid-dorsal hair-like setae on nymphal terga 1–9; and [40:2], the partial fusion of anterior sclerites of tergum 10 with the central sclerite, forming a narrow lateral cleft. Cardioperla nigrifrons (Kimmins, 1951) and Leptoperla varia Kimmins, 1951, formed a clade supported by [1:3], hook-like body setae in nymphs; and [32:2], complete fusion of the inferior branch of the M vein to the CuA vein in the adult hind wing (Fig. 4). Riekoperla perkinsi Theischinger, 1985, clustered with Senzilloides panguipullii (Navás, 1928), supported by [10:2], sparse fringe of hair-like setae on the extensor margin of the nymphal femur; [15:1], presence of a row of mid-dorsal processes or spines on nymphal terga 1–9; [40:2]; and [47:3], digitiform, straight adult epiproct with a wide base.

Finally, Notoperlopsis femina Illies, 1963, the sole representative of Zelandoperlinae included in this study, was consistently recovered as the sister group to Gripopteryginae. The synapomorphies supporting this clade are as follows: [10:0], absence of hair-like setae on the extensor margin of the femur; [16:0], presence of setiform, small, thin setae on the distal margin of terga 1–10; [36:1], the basal tarsal segment being half the length of the apical segment; [39:0], complete or partial (at least anterior half) fusion of the anterior sclerites of tergum 10; and [40:0], complete fusion of the anterior sclerites with the central sclerite of tergum 10.

The topology of all analyses remained identical across analyses with k values between 7–15 (Fig. 4). With the exception of Paragripopteryx, remaining genera included with more than one species appeared as monophyletic.

The main synapomorphies for the main clades are (for detailed synapomorphies for all clades see Fig. 4):

Clade B: Uncicauda testacea (Vera, 2006b) is sister to Limnoperla jaffueli (Navás, 1928), supported by [42:1], adult tergum 10 central sclerite with one denticle/tooth at the distal margin (Fig. 4).

Clade C: This large clade includes most Gripopteryginae, and is supported by [10:2], nymphal extensor margin of femur with a sparse fringe of hair-like setae; and [49:0], epiproct without a keel in the ventro-basal region.

Clade D: This clade forms a politomy, is sister of Teutoperla maulina and supported by four synapomorphies: [41:2], central sclerite of tergum 10, protruding, with lobe divided in two parts; [42:2], tergum 10, central sclerite, with two denticles; [47:2], epiproct digitiform, curved; and [50:1], epiproct inner surface with one row of denticles.

Clade E: This clade includes four genera and is supported by the following synapomorphies: [1:2], body with claviform-like setae; [11:1], femur extensor margin with small spine-like setae; and [16:3], distal margin of terga 1–10 with claviform-like setae. Within this grouping, Clade F, which includes Aubertoperla illiesi (Andean-Patagonian) and Gripopteryx species (Neotropical), is supported by [41:0], central sclerite of tergum 10 with a non-protruding distal margin. Clade H, comprising Rhitroperla rossi (Froehlich, 1960) (Andean-Patagonian) and Paragripopteryx species (Neotropical), is supported by [26:0], both wings unpigmented.

Clade J: This clade is supported by the following synapomorphies: [46:0], absence of a ventral sclerotized projection on the epiproct. It includes, in a polytomy, the Andean-Patagonian genera Andiperla, Andiperlodes, and Potamoperla, as well as the Neotropical genera Guaranyperla and Tupiperla, which together form a separate clade (L) (Fig. 4).

Clade L: Tupiperla species grouped with Guaranyperla species, supported by the following characters: [1:3], body setae hook-like; [8:0], nymphal metanotum mid-distal margin with a circular, concave shape; [17:2], distal margin of tergum 10 triangular; and [34:1], presence of a distal spine on the flexor margin of the femur (Fig. 4 and Figs S2–S3). Within this grouping, the analyses placed the new species in an unresolved polytomy with other Tupiperla species, while Guaranyperla species emerged as close relatives (Fig. 4). A detailed description is provided below.

Class Insecta Linnaeus, 1758

Order Plecoptera Burmeister, 1839

Family Gripopterygidae Enderlein, 1909

Subfamily Gripopteryginae Enderlein, 1909

Our findings revealed a polyphyletic Gripopteryginae, with Neopentura semifusca falling outside the clade, being more closely related to Antarctoperlinae taxa. In contrast, all remaining Gripopteryginae taxa were consistently recovered as a cohesive clade across all analyses (Fig. 4 and Figs S2–S3). These findings highlight the need for a re-evaluation of the taxonomic classification of Neopentura semifusca and bring light regarding the current arrangement of the Gripopteryginae subfamily.

The taxonomic history of Neopentura semifusca reflects its complex and controversial classification. Originally described by Illies (1965) within the family Penturoperlidae (currently Austroperlidae), the species was later reassigned by Illies (1969) to the family Gripopterygidae, specifically within Gripopteryginae (Vera 2006a). However, McLellan (1977), in his revision of Gripopterygidae subfamilies, excluded Neopentura from his classification. Subsequently, Vera (2006a), in his redescription of the genus, based on morphological characters, proposed a close relationship between Neopentura and the New Zealand genus Zelandobius (Antarctoperlinae); despite this, Neopentura was still categorized under Gripopteryginae by Froehlich (2010) in his Catalog of Neotropical Plecoptera.

Our results provide phylogenetic and additional morphological support for the closer relationship of Neopentura with Antarctoperlinae rather than with Gripopteryginae. The six synapomorphies provided in results, under Antarctoperlinae, placed Neopentura in a polytomy with Antarctoperlinae taxa (k = 9–15, Fig. 4) or in a clade with Pehuenioperla llaima (k = 7, Fig. S3).

This robust cladistic evidence, along with morphological evidence as previously suggested by Vera (2006a) and the characters defined by McLellan (1977), supports the hypothesis that Neopentura should be formally reclassified within Antarctoperlinae. This reclassification not only resolves the long-standing question regarding the placement of the genus but also strengthens the morphological coherence of Antarctoperlinae as a distinct clade within Gripopterygidae, as proposed by Pessacq et al. (2020). Therefore, Antarctoperlinae now includes the following genera: Antarctoperla, Araucanioperla, Ceratoperla, Chilenoperla, Ericiataperla, Megandiperla, Neopentura, Pehuenioperla, Pelurgoperla, and Plegoperla from South America, as well as Zelandobius from New Zealand.

Our analyses with k = 7–15, consistently recovered an identical Gripopteryginae clade (Clade A, Fig. 4) supported by four synapomorphies: the forewing CuA vein with one fork [31:1]; the sixth anal vein of the hind wing fused to the wing margin [33:1]; a simple, protruding lobe on the distal margin of the central sclerite in tergum 10 [41:1]; and the absence of a posterior sclerite in tergum 10 [43:0] (Fig. 4). None of these synapomorphies is exclusive, some are shared by other Gripopterygidae and some revert in some Gripopteryginae (see Fig. 4).

McLellan (1977, p. 120–121) proposed the absence of a posterior sclerite as a diagnostic character to synonymize the former subfamily Paragripopteryginae under Gripopteryginae and to establish Dinotoperlinae as a separate subfamily (some of its members also lack a posterior sclerite, e.g., Dinotoperla serricauda and Trinotoperla irrorata). The other diagnostic characters proposed by McLellan (1977) for Gripopteryginae and shared with other taxa were codified in our dataset (Table 1):

(i) [31:1], the CuA vein fork in the forewings, observed in most Gripopteryginae (except apterous species) and shared with some Dinotoperlinae taxa (e.g., Alfonsoperla flinti and Dinotoperla serricauda);

(ii) [33:1], hind wing’s sixth anal vein fused to the wing margin, identified as a synapomorphy for Gripopteryginae under k = 3 and k = 15, but also shared with some Australian Leptoperlinae taxa (e.g., Cardioperla nigrifrons, Newmanoperla thoreyi, and Riekoperla species);

(iii) [35:1], distoventral tibial spurs, observed in some Dinotoperlinae (e.g., Dinotoperla serricauda and Trinotoperla irrorata), and Leptoperlinae taxa (e.g., Cardioperla nigrifrons, Notoperla species, Riekoperla species, and Senzilloides panguipullii);

(iv) [in part 46:1], male genitalia with an unrecurved or absent epiproct tip, present in some Gripopteryginae genera.

In our analyses, four synapomorphies, including three characters used by McLellan (1977), consistently supported the monophyly of Gripopteryginae.

Our findings contrast with those of McCulloch et al. (2016), who, using molecular data (nuclear 18S, H3; mitochondrial COI), and including seven out of the fourteen Gripopteryginae genera, found no evidence supporting the monophyly of Gripopteryginae. Their results indicated significant polyphyly among Gripopteryginae taxa. For example, Tupiperla sp., Andiperla sp., and Aubertoperla sp. formed a clade with Australian Dinotoperlinae and Leptoperlinae. Similarly, Claudioperla tigrina was nested within the South American Antarctoperlinae. Other taxa, such as Gripopteryx sp., grouped with New Zealand Zelandoperlinae, while Limnoperla jaffueli and Teutoperla auberti clustered with Australian Leptoperla sp. (Leptoperlinae) (see fig. 3 in McCulloch et al. 2016).

Conversely, the morphological study by Pessacq et al. (2020) placed Gripopteryginae taxa within a clade containing Plegoperla punctata (Antarctoperlinae), a species with limited morphological detail available (Froehlich 1960; Illies 1963). However, their analyses focused solely on five Andean-Patagonian Gripopteryginae species and excluded Neotropical genera, limiting their conclusions.

Contrary to the conclusions of McCulloch et al. (2016) and Pessacq et al. (2020), our analyses, which include 13 out of the 14 currently recognized Gripopteryginae genera, provide morphological evidence supporting Gripopteryginae as a monophyletic clade, excluding Neopentura. These findings support the subfamily as established by Enderlein (1909) and later reinforced by McLellan’s (1977) review.

Although, in view of the conflicting results with McCulloch et al. (2016) and Pessacq et al. (2020), our own findings should be interpreted with caution, and they support a Gripopteryginae composed of the following taxa: Andiperla, Andiperlodes, Aubertoperla, Claudioperla, Falklandoperla, Gripopteryx, Guaranyperla, Limnoperla, Paragripopteryx, Potamoperla, Rhithroperla, Teutoperla, Tupiperla, and Uncicauda.

Regarding distribution, we found no clear biogeographic pattern: Neotropical and Andean-Patagonian taxa are intermixed, with the exception of the Neotropical genera Tupiperla and Guaranyperla, and the southern Andean taxa Andiperla and Andiperlodes, which nest together.

Duarte et al. (2022) highlighted that Paragripopteryx munoai did not cluster with other species of Paragripopteryx, suggesting it may represent a separate lineage. While the authors propose the potential need to establish a new genus and recognize Paragripopteryx as non-monophyletic in its current definition, they adopt a conservative approach, opting to await additional data before making taxonomic changes.

Morphologically, Paragripopteryx munoai differs from other Paragripopteryx species in several key characters: it lacks pterostigmatic crossveins in the forewings, has half-elliptical and shortened hind wings, and presents a W-shaped bilobed mid-distal margin on the metanotum in nymphs. Additionally, its femora and tibiae are bare, lacking the fringed setae typical of other species. The eggs are hemispherical and simple, in contrast to the elliptical eggs observed in related taxa.

Our study positioned Paragripopteryx munoai in an uncertain phylogenetic relationship relative to Paragripopteryx species analyzed, which suggests that the species may not belong within Paragripopteryx as currently circumscribed. As such, Paragripopteryx munoai represents a challenge, particularly due to its limited sampling and subtle morphological variation.

The collection of fresh specimens, especially from the type locality and surrounding regions, should be prioritize to enable molecular analyses and detailed comparisons with other Paragripopteryx species and closely related genera. High-resolution imaging of morphological structures, particularly male terminalia and egg morphology, may also aid in clarifying the boundaries of the genus. Additionally, expanded taxon sampling in phylogenetic analyses could help determine whether Paragripopteryx munoai forms an early-diverging lineage within Paragripopteryx, clusters with another genus, or warrants generic status on its own.

Our findings, along with those of McCulloch et al. (2016), reveal that the subfamilies Leptoperlinae and Dinotoperlinae are consistently polyphyletic, challenging their traditional classifications within Gripopterygidae. Additionally, Notoperlopsis femina (Zelandoperlinae), the sole representative of its subfamily in our study, showed a closer phylogenetic affinity with Gripopteryginae. However, due to limited sampling, its broader subfamily relationships remain inconclusive, no material from the remaining representatives of the subfamily was available to us.

We studied seven Leptoperlinae taxa (six out of the seven recognized genera). Our analyses clearly render the subfamily polyphyletic (Fig. 4). McCulloch et al. (2016) analyzed all Leptoperlinae genera and also recovered the subfamily as polyphyletic. In their study, Senzilloides panguipullii clustered with a group containing Dinotoperlinae taxa, Cardioperla (Leptoperlinae), and Tupiperla sp. (Gripopteryginae). Similarly, Pessacq et al. (2020) identified Senzilloides panguipullii as sister to a paraphyletic Gripopteryginae, although their study included only one Leptoperlinae species.

Our results also recovered Dinotoperlinae as polyphyletic and positioned at the base of the tree (Fig. 4). However, our sampling of only three of ten genera of the subfamily warrants a cautious interpretation. Although McLellan (1977) proposed the subfamily Dinotoperlinae based on members of Gripopteryginae and shares several characteristics with that family (Table 1), including the absence of the posterior sclerite, our findings indicate a distant relationship between them. These results partially contrast with those of McCulloch et al. (2016), who also recovered Dinotoperlinae as polyphyletic, analyzing 10 taxa from seven of the ten recognized genera. In their study, some dinotoperline taxa appeared closely related to Gripopteryginae, such as Dundundra sp. + Tupiperla sp. and Alfonsoperla flinti + Aubertoperla sp. Our analyses, however, left the relationships within the subfamily unresolved due to the inclusion of only three dinotoperline taxa.

One remarkable result of our study is that Trinotoperla irrorata consistently appeared in the first cladogenesis of Gripopterygidae, indicating an early divergence (Fig. 4). Similarly, Pessacq et al. (2020), who included five Dinotoperlinae taxa in their cladistic analysis, recovered Trinotoperla irrorata clustered with Vesicaperla kuscheli (Antarctoperlinae), further supporting a polyphyletic arrangement of Dinotoperlinae. However, Vesicaperla kuscheli was not included in our analysis.

Historically, Dinotoperlinae has been defined by a combination of morphological characters, including the absence of a posterior sclerite on tergum 10, the presence of a long CuA fork in the forewing, and the sixth anal vein of the hind wing free from the wing margin (McLellan 1977). However, taxa such as Alfonsoperla flinti, which possesses a posterior sclerite, challenge these traits as reliable synapomorphies, casting clear doubt on the monophyly of the subfamily.

The phylogenetic placement of taxa across these subfamilies highlights the need for revised classifications. Both Leptoperlinae and Dinotoperlinae may require redefinition or even division into more cohesive units. Achieving this will be necessary to expand samples of taxa to obtain a more comprehensive understanding of these subfamilies. In addition, the morphological characters traditionally used to define subfamilies should be complemented with molecular and biogeographic data to support robust phylogenetic analyses.