The Smicridea brasiliana species group includes five species distributed in northeastern Argentina and Brazil: Smicridea (Rhyacophylax) brasiliana (Ulmer), S. (R.) weidneri Flint, S. (R.) vermiculata Flint, S. (R.) arobasis Flint, and S. (R.) nanda Flint. The original descriptions of these species and their placement in the brasiliana species group were mainly based on the morphology of the male genitalia. However, the fine structure of the internal sclerites of the phallus, which proved to be useful for species delimitation, was not analyzed at the time. In this contribution, we provide a detailed description of the male genitalia and the morphology of the head, and analyze the shape of the wings using geometric morphometrics. The analyzed species can be easily differentiated by the shape of the phallus, especially by the structure of the internal sclerites, the shape of the head in dorsal view, and the shape of the cephalic setose warts. Furthermore, the geometric morphometric approach allowed their separation through the wing shape. The preliminary analysis of these features suggests that the brasiliana species group is not natural but its monophyly should be further tested within the framework of a phylogenetic analysis of all the species of the subgenus Rhyacophylax.

Forewing shape, male genitalia, Neotropical, Smicridea (Rhyacophylax) atrobasis, Smicridea (Rhyacophylax) nanda, Smicridea (Rhyacophylax) vermiculata, Smicridea (Rhyacophylax) weidneri

Smicridea is the only genus of Smicrideinae present in the Neotropical region (Schefter 1996; Flint et al. 1999). It is very diverse, represented by 255 described species grouped in two subgenera, Smicridea (Smicridea) McLachlan with 145 species and Smicridea (Rhyacophylax) Müller with 110 species (Holzenthal and Calor 2017; Alves et al. 2018; Mey and Ospina-Torres 2018; Sganga and Gibon 2018; Gibon and Sganga 2019; Rázuri-Gonzales and Armitage 2019; Vilarino et al. 2019; Desiderio et al. 2021; Queiroz et al. 2021; Santana et al. 2021). The taxonomy of Smicridea adult males has been studied extensively over the years, through the description of single species or the fauna of large geographic areas (e.g., Flint 1974). In the last decade approximately 74 new Smicridea species have been described (Albino et al. 2011; Rueda Martín and Sganga 2011; Oláh and Johanson 2012; Alves et al. 2018; Mey and Ospina-Torres 2018; Sganga and Gibon 2018; Gibon and Sganga 2019; Rázuri-Gonzales and Armitage 2019; Vilarino et al. 2019; Desiderio et al. 2021; Queiroz et al. 2021; Santana et al. 2021) but neither subgenus has been reviewed.

The Smicridea (Rhyacophylax) brasiliana (Ulmer, 1905) species group currently contains five species: S. brasiliana, S. weidneri Flint, 1972, S. vermiculata Flint, 1978, S. atrobasis Flint, 1983, and S. nanda Flint, 1983, that are distributed in northeastern Argentina and Brazil (Holzenthal and Calor 2017). This species group was established by Flint (1983) who did not provide a set of defining characters for the group but discussed the features that allowed the differentiation of these species (mainly the coloration, the presence of processes at the tip of the phallus, the shape of the internal sclerites, and the morphology of the tenth tergum and the inferior appendages). The relationships between these species were established by Flint in their original descriptions (Flint 1972, 1978, 1983). The identification of Smicridea species has long been based on the male genitalia, especially the structure of the tenth tergum and the phallus, allowing the delimitation of several species groups in both subgenera. These groups were never formally defined (following the principle of monophyly) but created to include species with similar characteristics. Because of this missing framework, several species in both subgenera were never placed in a species group. Oláh and Johanson (2012) summarized some of these groups, listed their defining characters, and the species included in them, in order to provide a framework to include their new species. However, a comprehensive work including a review and phylogenetic analysis of all Smicridea species, and a test of the validity of these species groups is lacking.

The morphology of the male genitalia has been extensively used for the delimitation of species in the order Trichoptera as a whole. Other characteristics of the adult morphology, such as the shape of the antennae and palps, the presence of ocelli, spur formula, shape and distribution of setose warts, and wing venation are usually used to identify families and genera (Holzenthal et al. 2007). In the genus Smicridea the structure of the internal sclerites of the phallus has been proven to be useful for the differentiation of species in the subgenus Rhyacophylax, mainly in closely related ones (e.g., Rocha et al. 2016; Santana et al. 2021). Despite its importance, this character has not been analyzed in detail in the early descriptions of Smicridea species, which is the case for S. brasiliana, S. nanda, and S. weidneri. A few authors have analyzed non-genital characters for the identification of Smicridea species. Oláh and Johanson (2012) and Sganga and Gibon (2018) used the maxillary palp formula to represent the length ratio of the 5 palp articles. More recently, Vilarino et al. (2019) explored the use of new characters to evaluate species delimitation such as the presence and shape of head setose warts and sutures, eye size, and forewing forks. All these features were variable among the species described and represent a source of characters for future phylogenetic analysis.

In recent decades, the number of geometric morphometric studies in insects has increased in the literature. This methodology became a powerful tool to detect minimal shape variations which often are undetectable by traditional morphological studies and emphasizes differences between groups (Villemant et al. 2007). It is usually applied to distinguish species (Baylac et al. 2003; Lorenz et al. 2017; Simões et al. 2020), identify population structure (Kiyoshi and Hikida 2012; Kamimura et al. 2020), sexual dimorphism (Gushki et al. 2018), study morphological evolution during ontogeny (Springolo et al. 2021), and map phylogenetic hypotheses (Huang et al. 2020), among others. Studies based on the taxonomic delimitation of species that are difficult to solve by traditional anatomical methods have been carried out in various insect taxa (Sábio et al. 2014), many of them based on wing geometry (Kiyoshi and Hikida 2012; Shimabukuro et al. 2016; Huang et al 2020; Simões et al. 2020). The wings of the species of Smicridea (Rhyacophylax) are very conserved in the arrangement of their veins, which makes them an excellent material to investigate interspecific variations for the delimitation of species.

In the present work, we aimed to redescribe the species in the Smicridea brasiliana group offering a detailed description of the genital segments, especially the phallus, and provide non-genital characters for their identification. Additionally, we tested the use of geometric morphometrics for species delimitation. Integrative taxonomic investigations, which include traditional tools together with modern methodologies, are increasingly being implemented to solve species delimitation problems (González et al. 2019). Geometric morphometrics techniques show high performance in this task (Mutanen and Pretorius 2007). This work represents the first study that incorporates the geometric morphometric approach to the taxonomy of the order Trichoptera, in particular the genus Smicridea.

Specimens of Smicridea brasiliana, S. nanda and S. weidneri housed in the National Museum of Natural History, Smithsonian Institution, Washington DC (USNM) were examined. Those specimens were identified and loaned by Dr. Oliver Flint Jr. Additionally, we borrowed specimens of S. atrobasis and S. (R.) vekona from the Facultad de Humanidades y Ciencias, Universidad de la República (Uruguay, FHCM) and the Instituto de Biodiversidad Neotropical, CONICET-Universidad Nacional de Tucumán, (Argentina), respectively. Specimens of the other species treated herein were collected in Salto Encantado Provincial Park and Forest Refuge and research center Antonia Ramos (Misiones province, Argentina).

The samples were collected in December 2004 and November 2013 using light and Malaise traps. The specimens obtained were fixed and preserved in 80% EtOH. Voucher specimens were deposited at the Museo Argentino de Ciencias Naturales (Buenos Aires, Argentina).

For identification and illustration of the specimens the abdomen was cleared using a hot 10% NaOH solution. Then the cuticle was rinsed in distilled water, neutralized with acetic acid, and mounted in a dish with glycerin for observation. Line drawings of the genital structures were produced using a camera lucida attached to a microscope. Line illustrations of the heads were constructed using photographs as templates, which were obtained with a digital camera fixed to a stereomicroscope. All the images were digitalized with Adobe Illustrator (v. 15.0.0 Adobe Systems Inc.).

For the description of the heads the following distances were measured (Fig. 1):

Figure 1. 

Heads in dorsal view of the species of the brasiliana group A Smicridea (Rhyacophylax) brasiliana B S. (R.) weidneri C S. (R.) vermiculata D S. (R.) atrobasis E S. (R.) nanda. Abbreviations: CSL length of the coronal suture, IOD interocular distance, MEW maximum eye width, MHW maximum head width. Scale bar: 1 mm.

IOD interocular distance;

CSL length of the coronal suture;

MEW maximum eye width;

MHW maximum head width.

All the measures were taken using a stereomicroscope with a graduated eyepiece. The terminology used by Albino et al. (2011) was followed for the description of the male genitalia, the one from Wells and Neboiss (2018) for the setose warts, and the one from Oláh and Johanson (2007) for the cranial areas.

For the morphometric analysis, all the species included in the brasiliana group were used along with five additional species from the same subgenus, in order to increase the discriminatory power of the methodology.

The left forewings of males (n = 154) of Smicridea (Rhyacophylax) mesembrina (Navás, 1918) (n = 21), S. weidneri (n = 16), S. vermiculata (n = 22), S. (R.) spinulosa Flint, 1972 (n = 18), S. atrobasis (n = 18), S. (R.) vekona Oláh & Johanson, 2012 (n = 19), S. (R.) pampeana Flint, 1980 (n = 18), S. (R.) unguiculata Flint, 1983 (n = 20), S. nanda (n = 1), and S. brasiliana (n = 1) were dissected. Then, the removed wings were extended and mounted on a slide, using alcohol as medium, and covered with a coverslip. The alcohol was left to evaporate before taking photographs with a digital camera fixed to a stereomicroscope (two photographs were taken of each wing). Cartesian coordinates of ten landmarks of each wing (Fig. 2) were digitized using tps-UTILS v. 1.38 (Rohlf 2006a) and tps-DIG v. 2.05 (Rohlf 2006b). The landmark configurations were scaled, translated and rotated using the GLS Procrustes superimposition method (Bookstein 1991) using the MorphoJ software v. 1.06d (Klingenberg 2011) and subsequently a thin-plate spline analysis was performed allowing the visualization of shape differences as deformation.

Figure 2. 

Forewing of Smicridea (Rhyacophylax) mesembrina showing the location of the selected landmarks (1–10).

Permutation tests for distances by species pairs (20,000 rounds of permutations) for the Mahalanobis (Table 1) and Procrustes distances were performed. Then the percentage of correct reclassification by pairs of species was calculated from the cross-validation procedure.

Canonical variate analysis (CVA) was performed on aligned landmark coordinates and the specimens were reclassified to each species (jackknife method) to evaluate the effectivity of the discriminant analysis for assigning them to their own group using the software Past v. 4.02 (Hammer et al. 2001).

The species S. nanda and S. brasiliana were excluded from all the statistical tests mentioned above due to an insufficient number of specimens.

A principal component analysis (PCA) with the consensus configurations of the species of the brasiliana group (S. brasiliana, S. weidneri, S. vermiculata, S. atrobasis, and S. nanda) was performed. In addition, the thin plate spline method was used to illustrate the transformations of the wing shapes compared to the consensus wing shape of the group. Mahalanobis distances between the mean shapes of each species of the brasiliana group were used to construct a dendrogram using the Unweighted pair-group method with arithmetic mean (UPGMA) with the software Past v. 4.02 (Hammer et al. 2001).

For wing size analysis, centroid size (CS) was used as a measure of size and was computed as the square root of the sum of squared distances from all landmarks to the centroid of the landmarks configuration (Bookstein 1991). The CS variation for each species is shown with a violin-plot. Differences in CS among species were assessed through a Kruskal-Wallis test and a posteriori pairwise test. Smicridea nanda and S. brasiliana were not included in this analysis due to insufficient number of specimens, but the CS of both species are shown in the plot.

Based on the analysis of the fine structure of the genitalia of the species of the brasiliana group we consider that the group is heterogeneous. Smicridea brasiliana, S. weidneri, and S. nanda seem to be closely related species with complex phallic structures, including internal plates of different shapes associated with the distal end of the sclerotized ejaculatory duct that are absent in S. atrobasis and S. vermiculata. In particular the genitalia of S. vermiculata is most similar to that of S. dentifera Flint and S. unguiculata Flint, which are unplaced to species groups, that have a simple ejaculatory duct and lateral subapical spines at the phallus. The analysis of the morphology of the heads of these species also supports a closer relationship between S. brasiliana, S. weidneri, and S. nanda, with anterolateral setose warts bifid and posterior setose warts triangular. The head of S. vermiculata is more similar in shape to these species but differs in the structure of the posterior setose warts that bear internal lobes and the postgenal areas that are more developed. In contrast, the head of S. atrobasis is unique in the group, with a more quadrangular outline, very large eyes, a reduced interocular area, and oval anterolateral setose warts. The genitalia of this species is also different from the rest: the apex of the ejaculatory duct is shaped like an arrowhead in dorsal view and the phallus ends in two laterodorsal and two lateroventral lobes. The presence of apicolateral lobes on the phallus is also found in S. (R.) mesembrina, not placed in a species group. The relationships between S. brasiliana, S. weidneri, and S. nanda are also strengthened by the configurations of the forewings, as was observed through the geometric morphometrics analysis.

The placement of S. atrobasis in the brasiliana group is conflictive. As stated before, although the configuration of the forewing of this species is similar to that of S. weidneri, the genitalia and the morphology of the head of both species differ. The relationships of the forewing configurations in the brasiliana species group were compared in the dendrogram, with and without S. atrobasis. The exclusion of this species from the analysis shows the same patterns that the ones observed using the morphology of the genitalia and features of the head. In this context, the similarities in the forewing configuration could be seen as a homoplasy rather than a homology, although further phylogenetic analyses are needed for confirmation.

The geometric morphometric analysis of wing shapes was useful for discriminating the species herein studied. This is the first study that uses this methodology in the order Trichoptera and needs to be examined in more species of Smicridea, and other caddisfly taxa as well, using not only wing shapes but other structures of the body. The larvae of Smicridea (Rhyacophylax) are good candidates to test this approach. In this subgenus the larvae are generally very similar, with no clear defining characters to separate them, but there are subtle interspecific differences in the shape of the head and the frontoclypeal apotome (JS pers. obs.). These differences in shape could be tested with this methodology. Taking into consideration that landmark configurations can be used in phylogenetic reconstructions (Catalano et al. 2010, 2015; Palci and Lee 2019), the exploration of this type of characters in the study of Trichoptera can be of great relevance.

In this study, we provided a new approach for the delimitation of species in the genus. The head morphology is somewhat overlooked in the descriptions of most Trichoptera species. Here, we propose a more comprehensive approach including more detailed descriptions of relevant characters, besides the male genitalia, that would be useful for differentiating closely related species. Likewise, we demonstrated that the geometric morphometrics analysis of wing shapes can be used to discriminate the species of Smicridea (Rhyacophylax) herein studied. This fast, simple, and inexpensive method proved to be an efficient technique to confirm the identity of the specimens and could potentially be used to differentiate cryptic species, which were previously reported in Smicridea and other insect genera (Pauls et al. 2010; Chroni et al. 2018; Chatpiyaphat et al. 2021). Furthermore, it can also be a source of characters for phylogenetic analysis, not as a substitute for traditional morphological characters, but rather as a complementary descriptor of shape diversity (Palci and Lee 2019).

Even though the analyzed features indicate that the brasiliana group might not be a natural group as informally defined, the relationships between these species and the rest of the species in the subgenus Rhyacophylax must be established by a phylogenetic analysis and the monophyly of all the current groups of species should be tested.

We appreciate the comments and suggestions made by the two reviewers that improved the quality of the manuscript. We specially thank the late Dr. Oliver Flint, Jr. (National Museum of Natural History, Washington DC, USA), Dr. Enrique Morelli (Facultad de Humanidades y Ciencias, Universidad de la República, Uruguay) and Dr. Paola A. Rueda Martín (Instituto de Biodiversidad Neotropical, CONICET-Universidad Nacional de Tucumán, Argentina) for the loan of specimens. We extend our gratitude to the forest rangers of Salto Encantado Provincial Park and Forest Refuge and research center Antonia Ramos for field assistance and to Alejandro A. Fernandez for laboratory assistance. We also thank the Ministerio de Ecología y Recursos Naturales Renovables from Misiones Province for its support and for providing the permits to sample in the protected areas.

This project received financial support from UBACyT 20020130200240BA and Departamento de Biodiversidad y Biología Experimental, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires.