Introduction

The knowledge of scorpion diversity in North America has improved recently (^{González-Santillán & Prendini, 2013}; ^{Goodman, Prendini, Francke et al., 2021}; ^{Ponce-Saavedra & Francke, 2019}; ^{Santibáñez-López et al., 2014}); however, much remains to be discovered. Local-scale inventories may be a solution to unveil species communities that, in turn, can help conform regional faunas. This approach has rarely been applied to study scorpion diversity. Furthermore, few local faunal studies have been conducted in Mexico, and only a handful of them have been published; other revisionary contributions included limited fieldwork effort (^{Baldazo-Monsivais et al., 2012}, ²⁰¹⁶, ²⁰¹⁷).

The International Barcode of Life (iBOL) has grown as a powerful tool for discovering biodiversity, among other applications (https://ibol.org). Scorpion barcoding studies have permitted the identification and delimitation of species in several regions of the world (^{Fet et al., 2014}, ²⁰¹⁶; ^{Goodman, Prendini, & Esposito, 2021}; ^{Podnar et al., 2021}). Despite the high diversity of scorpions in Mexico -an update by ^{Ponce-Saavedra et al. (2023)} comprises 311 scorpion species- only 1 mini-barcoding study has been conducted (^{Goodman, Prendini, & Esposito, 2021}). Herein, we present a second scorpion barcoding study for this country but aim at discovering the components of a local scorpion assembly. Colima exhibits a complex topology comprising littorals with a tropical climate and extreme topological variation from sea level to mountain ranges rising to over 4,000 m in approximately 5,600 km². Colima’s territory supports a rich local flora and fauna (^{Ramírez-Ruiz & Bretón-González, 2016}). Colima lies between the limits of the Nearctic and Neotropical biogeographical Realms (Fig. 1B). Beyond its complexity, Colima represents an enclosed littoral surrounded by mountain ranges and 3 large rivers that divide the territory into 2 sections (Fig. 1), which makes it a well-defined and manageable geographical unit ideal for studying a unique community of scorpions. Thus far, 2 families, 3 subfamilies, 5 genera, and 12 species of scorpions have been identified in Colima (Table 1). The knowledge of scorpion diversity and evolutionary studies in Mexico has steadily been unveiling one of the vastest biodiversity hotspots in the world. For instance, 2 of the most diverse scorpion families and subfamilies, Vaejovidae (Syntropinae) and Buthidae (Centruroidinae), have been treated in recent phylogenetic and taxonomic analyses (^{Esposito & Prendini, 2019}; ^{González-Santillán & Prendini, 2013}, ²⁰¹⁵).

Figure 1 Map of the west coast of Mexico. A, Orographic and hydrographic elements of Colima (COL) and the surrounding states of Jalisco (JAL) and Michoacán (MIC). 1, Manantlán Sierra; 2, massive Cerro Grande; 3, Colima Volcano; 4, Marabasco or Cihuatlán River; 5, Armería River; 6, Coahuayana River. B, Biogeographical provinces (^{Morrone et al., 2017}). Area within the green line corresponds to the Sierra Madre del Sur and north Colima Volcano Trans Mexican Volcanic Belt province (^{Morrone et al., 2017}) -notice that both provinces are connected in Colima. Area outside the green line corresponds to the Pacific Lowlands province (^{Morrone et al., 2017}). Orographic components are indicated in gray scale from light low elevation to dark high elevation. 

This contribution aims to survey the scorpion richness in Colima, using not only the COI-barcoding genetic marker but 2 additional mitochondrial markers and 1 nuclear marker to establish a framework to build a stable and predictable taxonomy. Taking advantage of robust phylogenies produced for the 2 families distributed in Colima, we use these topologies as a baseline for comparison to test the presence of several previously reported species and to taxonomically circumscribe our fresh samples. Unlike other barcoding studies, our approach seeks to unveil the richness within the state instead of focusing on delimiting the species of a taxonomic group of scorpions.

Materials and methods

We conducted field collections during May and September 2015-2018 in various ecosystems, including tropical deciduous forest, oak-pine, and tropical forest within the state of Colima, at elevations ranging from 47 to 2,200 m. (Table 2). Logistically, we leveraged our collection sites with the help of private landowners who gave us access to their property. Specimen collection methods included direct collection during the day by moving objects on the ground or by ultraviolet detection during the night. To preserve specimens, ethyl alcohol at 90% was used and stored at -80 °C. Each specimen lot carried a label with coordinates and locality information. We obtained scorpions from 14 localities and sequenced 18 samples of Centruroides from Colima (Table 2). We also included 8 species from outside the state to test the presence of C. elegans, C. infamatus, and C. limpidus reported previously in the literature (^{Ponce-Saavedra et al., 2016}); C. noxius and C. suffusus for comparative purposes and samples of C. tecomanus and C. ornatus from GenBank with a total of 31 specimens of the genus Centruroides within 4 species groups. Unlike buthids, we obtained 12 samples of vaejovids to include in this analysis. To evaluate the identity of vaejovids, we used the BLAST® suite (https://blast.ncbi.nlm.nih.gov/Blast.cgi) to search for similar sequences deposited in the nucleotide collection database at NCBI. Thorellius has been revised recently (^{González-Santillán & Prendini, 2018}). Therefore, several DNA sequences are available, and fewer sequences of Mesomexovis sp. and Vaejovis sp. were available in GenBank, as they are still unrevised. Using the genetic markers as queries, we obtained 10 additional samples. The total number of taxonomic specimens used for these analyses was 53 (Tables 2, 4).

Genomic DNA was extracted from the legs or pedipalp of specimens using Qiagen Dneasy/trisol method Tissue Kits or a DNAzol Genomic DNA isolation Reagent kit (Molecular Research Center INC, Cincinnati, Oh). We amplified 3 mitochondrial markers, 12S rDNA, 16S rDNA, and the barcode COI, and the nuclear marker 28S rDNA. We performed the Polymerase Chain Reaction with the following thermal profile: an initial denaturation step (3 min at 94°C) followed by 35 cycles including denaturation at 94°C for 30 s, annealing (46-55°C) for 30 s, and extension at 72°C for 30 s, with a final extension step at 72°C for 7 min. The PCR reaction was conducted using PureTaq-Ready-To-Go PCR Beads (GE Healthcare), 2 µl of DNA template, 21 µl of DNA grade H₂O, and 1 µl of each direction primer listed in Table 3. We verified PCR products with a 1% agarose-TBE electrophoresis gel stained with CYBR Safe. For purification of the amplified products, we used Ampure Magnetic Beads (Beckman-Coulter) and re-suspended in 40 µl DNA grade water by using a Beckman Coulter Biomek NX 18 robot. Each 8 µl cycle-sequenced reaction mixture included 1 µl of Big Dye, 1 µl of Big Dye Terminating buffer, 1 µl of 3.2 pm primer, and 5 µl of gene amplification product. Cycle-sequenced products were purified with CleanSeq magnetic beads on a Biomex NX robot. Products were re-suspended in EDTA, and 33 µl were processed in an Applied Biosystems, Inc. Prism 3730xl automated DNA sequencer. These products were sequenced with the same primer pairs used for amplification at the Laboratorio de Secuenciación Genómica de la Biodiversidad, at Instituto de Biología and Unidad de Síntesis y Secuenciación de DNA, Instituto de Biotecnología, UNAM. The sequences were edited using Sequencher® version 5.4.6.

Each genetic fragment was aligned separately for all terminals with MAFFT using the online server (https://mafft.cbrc.jp/alignment/server/). Since the number of nucleotides per gene was similar, we used the G-ins-i iterative refinement method, as recommended elsewhere (^{Katoh et al., 2019}; ^{Kuraku et al., 2013}), and other parameters were kept default. To select the best fit of the substitution model per partition and conduct the phylogenetic analyses, we used IQ-TREE version 2 (^{Kalyaanamoorthy et al., 2017}; ^{Nguyen et al., 2015}) and we estimated branch support with 1,000 replicates of the ultrafast bootstrap (UFBOOT) algorithm (^{Hoang et al., 2018}). Furthermore, each genetic marker was analyzed individually to explore its phylogenetic signal and contribution to the final topology. We conducted concatenated and partitioned analyses, handling all matrices in Mesquite (^{Maddison & Maddison, 2023}). For the COI partition we explored the best codon partition per site, but the results had no effect on the topology. Additionally, to evaluate each marker and nucleotide site within each marker, we calculated the gene (GCF) and site (SCF) concordance factors on the topology that we emphasized in the discussion (^{Mihn et al., 2020}).

Distributional records and maps. We obtained records with geographical coordinates of the scorpion species treated here via the Global Biodiversity Information Facility (GBIF) from the InDRE, responsible in Mexico for epidemiological vigilance (^{Huerta-Jiménez, 2018}), and the records published by ^{Ponce-Saavedra et al. (2015)}. These records were the basis for the species distribution maps depicted in Figures 3 to 6. We used the program QGIS 3.16.6-Hannove (^{QGIS, 2021}) to create the distributional maps. The topological model with the data was from ^{Jarvis et al. (2008)}, and to draw the political boundaries we used shapefiles obtained from Conabio. The biogeographic regionalization of Mexico into provinces and districts follows ^{Morrone et al. (2017)} and ^{Morrone (2019)}.

Figure 2 Phylogenetic tree from an analysis of 4 concatenated genetic markers, 3 mitochondrial (12S, 16S, and COI) and 1 nuclear (28S). The topology shows families, subfamilies, and species groups. Numbers beside nodes indicate ultrafast bootstrap/genetic concordance factor/site concordance factor. Colored species represent the 11 species found in our collection within the study area and samples grouping with them distributed inside or outside Colima. Scorpion photos: upper right, Centruroides tecomanus from El Palapo, adult male; middle, Centruroides hirsutipalpus from Sierra de Minatitlán, adult female; lower left, Thorellius intrepidus from El Palapo, adult male. 

Figure 3 Map of the west coast of Mexico with georeferenced filtered records from the InDRE database. A, Distribution of Centruroides elegans (circles), records north of Sierra de Manantlán may be misidentifications; B, distribution of Centruroides tecomanus (circles), records in Guerrero (GRO), Guanajuato (GTO), Jalisco (JAL), and Nayarit (NAY) may be misidentifications; Centruroides tecomanus 1 green cross, and Centruroides tecomanus 2 pink crosses. State abbreviations: AUG, Aguascalientes; COL, Colima; MEX, Estado de México; QRO, Querétaro; SLP, San Luis Potosí; ZAC, Zacatecas. 

Figure 4 Map of the west coast of Mexico with georeferenced filtered records from the InDRE database and for B additional records from ^{Ponce- Saavedra et al. (2015)}. A, Distribution of Centruroides infamatus (circles); B, distribution of Centruroides ornatus (circles), our sample in Colima (cross). 

Figure 5 Map of the west coast of Mexico with georeferenced filtered records from the InDRE database. A, Distribution of Centruroides hirsutipalpus (red cross), Centruroides sp. 1 (blue cross), Centruroides sp. 2 (pink cross), and Centruroides possanii (green crosses); B, distribution of Mesomexovis aff. occidentalis sequences in this work and a sample from Chamela Jalisco recorded by ^{González-Santillán (2004)}, conspecific with our samples in green crosses, Vaejovis sp. (mexicanus group) (pink cross). 

Figure 6 Map of the west coast of Mexico with georeferenced filtered records from the InDRE database. A, Distribution of Thorellius intrepidus green squares, our samples in Colima pink crosses; B, distribution of Thorellius cristimanus green squares, our samples in Colima pink crosses. 

Results

The gene fragments that we obtained are listed in Table 4 and the main statistics of the alignment and concatenated matrix are in Table 5. Our topology produced a clade representing members of the family Buthidae and another clade representing Vaejovidae (Fig. 2). Within buthids, the first clade included C. huichol and C. noxius, component species of the bertholdii species group (^{Ponce-Saavedra & Francke, 2019}), supported by 95% UFBOOT, 100% GCF, and 40% SCF. The next clade included C. elegans, C. limpidus, and a putative undescribed species with lower support values of 68%, 100%, and 38%, respectively, including members of the elegans group. Although with low support, the bulk of species appeared in the third clade comprising species within the infamatus group, with C. tecomanus, C. infamatus, C. possanii, C. hirsuticauda, C. ornatus, and 2 putative new species. The last clade included C. suffusus and C. sculpturatus, which ^{Ponce-Saavedra and Francke (2019)} circumscribed within the infamatus and the elegans species group, respectively (Fig. 2). However, the overall topology retrieved herein is concordant with the North American clade of the genus Centruroides (^{Esposito & Prendini, 2019}).

The Vaejovid clade received full support, except SCF (60%), which included 3 genera: Vaejovis, Mesomexovis, and Thorellius (Fig. 2). The genera occurred in a topology that overall resembles that of ^{González-Santillán and Prendini (2013}, ²⁰¹⁵⁾; while Vaejovis is a genus within Vaejovinae, Mesomexovis and Thorellius are part of the Syntropinae subfamily. Within Syntropinae, only Thorellius was monophyletic, and T. intrepidus received full support. Additionally, we identified 2 putative new species belonging to the genera Vaejovis and Mesomexovis.

Distributional maps and species geographical ranges. We mined 6,965 records of Mexican scorpions from the InDRE database but retained only 1,000 of the species treated herein. We filtered the records using previously published identified species by specialists and the distribution of the species cited for Colima (^{González-Santillán & Prendini, 2013}, ²⁰¹⁸; ^{Lourenço & Sissom, 2000}; ^{Ponce-Saavedra et al., 2016}; ^{Sissom, 2000}).

The distribution of Centruroides species appears to be restricted by physiographical elements. Centruroides elegans and C. tecomanus are restricted to the coastal lands of Jalisco and Colima, respectively (Fig. 3). The Sierra Madre del Sur appears to be the main barrier to the north of their distribution. The records of these species overlap broadly with a visible gap created by the Sierra de Minatitlán (Fig. 3B) either by subsampling or by an effective geographic barrier.

Centruroides ornatus and C. infamatus, on the other hand, are restricted to the Transmexican Volcanic Belt (TVB) (Fig. 4). While the former species has an entire distribution within this province (Fig. 4B), the latter seems to be distributed in patches, one to the north, along the border of the Chihuahuan desert and the TVB, and the second to the south, between the Balsas Depression and the TVB (Fig. 4A).

Centruroides possanii appears to be a microendemic scorpion species component of the fauna restricted to a massive karstic lone mountain, Cerro Grande (^{González-Santillán et al., 2019}). Two putative undescribed species of Centruroides occupy the extreme east and west of the coastal line of Colima, and C. hirsutipalpus appears restricted to the Sierra Minatitlán (Fig. 5A).

Of the vaejovid species, thus far, the only records for Mesomexovis sp. are in Colima, albeit there is one conspecific record in Chamela, Jalisco (^{González-Santillán, 2004}). The putatively undescribed species of Vaejovis appears restricted to Cerro Grande (Fig. 5B), like Centruroides possanii, potentially another microendemic species. Thorellius intrepidus and T. cristimanus are widely distributed within Colima and across the TVB, the Balsas Depression, and the Sierra Madre del Sur (Fig. 6).

Discussion

Of the 2 families found in Colima, Buthidae comprise 7 species of Centruroides, becoming the most diverse (Table 1, Fig. 2). Our multigene analyses suggest that the previously recorded species C. elegans, C. infamatus, and C. limpidus reported by ^{Ponce-Saavedra et al. (2016)} might not be part of the scorpion fauna of Colima, as we demonstrate in the following sections.

Centruroides elegans and C. limpidus were not found in Colima. Centruroides elegans has an obscure taxonomic history. Firstly, its original description is too general and never indicated a precise type locality, but “Mexico” (^{Fet Lowe, 2000}). Secondly, although some taxonomic works clarified its former subspecies, the nominal taxon identity of C. elegans remains ambiguous. While ^{Lourenço and Sissom (2000)} suggested that this species is distributed in Jalisco, later, ^{González-Santillán (2004)} concluded that its distributional limits have never been defined with precision. The 2 exemplars of C. elegans collected in Chamela, Jalisco, grouped with members of the elegans group, sisters to C. limpidus from Iguala, Guerrero, and these in turn, were sister to our 2 exemplars identified as Centruroides new sp. 2 11 and 14 from La Central, in the municipality of Manzanillo (Figs. 2). The distributional data from InDRE of C. elegans is limited right at the northern border of Colima, except for 9 records away from that geographic barrier (Fig. 3A), which we hypothesized as a misidentification due to their position outside the Pacific coastline. Our fieldwork produced no sample conspecific to C. elegans but did produce other genetically distant exemplars. Until denser sampling along the northern border of Colima is conducted, we conclude that C. elegans and by corollary C. limpidus might not be part of the scorpion fauna distributed in Colima as suggested by ^{Ponce-Saavedra and Francke (2013)} and ^{Ponce-Saavedra et al. (2016)}.

Centruroides infamatus may not be part of the Colima scorpion fauna. Centruroides infamatus is another obscure taxon with an ambiguous distributional pattern from published data. Despite its inclusion in phylogenetic analyses (^{Quijano-Ravel et al., 2019}; ^{Towler et al., 2001}), its taxonomic circumscription and distribution have never been clarified. Among other problems, the original description indicates “type locality unknown” (^{Fet & Lowe, 2000}). ^{Hoffmann (1932)} studied the scorpions from Mexico and realized that the specimens from Michoacán agreed with the original description of the species and proposed that its distribution was expanded to Central Mexico from the Pacific Coast in Sinaloa and Colima to Guanajuato. We included 2 exemplars of C. infamatus, 1 from León, Guanajuato (C. infamatus 15 scrp), and 1 terminal from Tandamangapio municipality, Michoacán (C. infamatus LP1822; Esposito et al., 2018). However, the terminal LP1822 grouped with exemplars of C. ornatus (Fig. 2) and Tandamangapio is approximately 16 km from Sahuayo and Jiquilpan, 2 localities recorded in the redescription of C. ornatus (^{Ponce-Saavedra et al., 2015}).

We plotted 612 records from the InDRE database of C. infamatus to compare the distribution of C. ornatus (Fig. 4A). Although northern Michoacán is almost exclusively occupied by C. ornatus, Jalisco and Michoacán present an overlap with C. infamatus (Fig. 4). Additionally, the InDRE database contains records of C. infamatus from Oaxaca, Puebla, Nayarit, Durango, and Sinaloa (not shown), most likely occupied by other species, implying that most records are misidentifications. The species that morphologically could be confused with C. infamatus in the states of Oaxaca and Puebla are C. baergi, C. nigrovariatus, or C. rodolfoi because the overall base color, body size and carapace pigmentation are similar among the species. The same morphological features are present in specimens from Nayarit and Durango, however, the identity of the populations in these two states have never been studied with a molecular approach and may represent a geographically distinct species.

Nevertheless, recently ^{Ponce-Saavedra et al. (2022)} described C. baldazoi for Sinaloa, related morphologically with C. infamatus, potentially a species with which it can be mistaken. Although Ponce-Saavedra and collaborators included C. infamatus in the distribution of Sinaloa and Colima, they failed to indicate precise localities. In summary, the distribution of C. infamatus remains unsolved, attested by the wide distribution reported in the InDRE database. Thus, the infamatus species complex requires comprehensive molecular analyses to delimit its taxonomic circumscription and geographical distribution. On the other hand, our samples of C. ornatus 3 and 4 grouped with C. ornatus LP1822 and with C. ornatus 2003 from Morelia, Michoacán (^{Esposito & Prendini, 2019}), forming a monophyletic group (Fig. 2). Since the samples LP1822 and 2003 lie within the area of C. ornatus proposed by ^{Ponce-Saavedra et al. (2015)}, we concluded that our exemplars are conspecific with C. ornatus and expanded the distribution southwards, drawing a distributional border in northern Colima (Fig. 4B).

In conclusion, we propose that C. infamatus might not be present in Colima, but further sampling and analyses are needed. Despite the morphological similarity between C. infamatus and C. ornatus, our molecular analysis indicates genetic differences and may present a distinctive distributional pattern yet to be drawn with more samples.

Centruroides tecomanus is a species complex. Due to their morphological similarity with C. limpidus, C. tecomanus was described as C. limpidus tecomanus (^{Hoffmann, 1932}), and the author delimited C. tecomanus distribution to the lowlands of Colima and surmised that its distribution extended to the south along the coastline of Michoacán. But, most importantly, Hoffmann mentioned that on the northern coastline, C. elegans substitutes C. tecomanus. In contrast, the distribution recorded in the InDRE database for C. tecomanus implies that these 2 species present a wide area of sympatry on the coast of Jalisco (Fig. 3).

^{Ponce-Saavedra et al. (2009)} proposed C. tecomanus as a “bona fide species” and assumed its distribution includes the coastline of Michoacán, following ^{Hoffmann (1932)}. In their molecular and morphological analyses, the authors never included exemplars from Tecomán, Colima, the type locality for the species, assuming that the specimens from Michoacán were conspecific with the populations of Tecomán. Furthermore, ^{Quijano-Ravell et al. (2010)} extended the distribution of C. tecomanus to 4 localities in Guerrero and a similar number in Jalisco in a montane area.

The phylogenetic tree presented here substantiates 2 clades of C. tecomanus within the monophyletic infamatus species group (Fig. 2). We hypothesize that these clades may represent potentially distinct sympatric species, emphasizing that our topology obtained full support for these clades with multiple samples. Centruroides tecomanus2 appears to be more common and widely distributed in Colima, whereas C. tecomanus1 is less common, with only 2 localities grouped with Centruroides tecomanus 2007 from Michoacán and from the municipality of Comala, Colima, Centruroides tecomanus (Fig. 3B). Our results suggest that the populations distributed in Michoacán may represent a cryptic, undescribed species; consequently, the exemplars reported in Guerrero (^{Quijano-Ravell et al., 2010}) are unlikely to be conspecific with C. tecomanus. The authors’ discovery of new populations in Jalisco and Guerrero were based entirely on their analyses of morphological characters, precisely the most common way of confusing cryptic species.

Considering that the montane border between Colima and Michoacán is occupied by Centruroides romeroi Quijano-Ravell, de Armas, Francke, ^{Ponce-Saavedra, 2019}, it is fair to assume that C. tecomanus1 inhabits coastal ranges following the coastline of Michoacán to the Lázaro Cárdenas delta, where it may be substituted by Centruroides bonito Quijano-Ravell, Teruel, Ponce-Saavedra, 2016 or even other undescribed species. The Balsas River has been proposed to be a geographical barrier for several epigean arachnids, such as Amblypygi and Theraphosidae (^{Mendoza & Francke, 2017}; ^{Schramm et al., 2021}) and for small mammals (^{Ruiz-Vega et al., 2018})

Noteworthy is the locality La Central, right on the border between Colima and Jalisco, a few kilometers southeast of the Marabasco River (Fig. 1A), where we collected 3 putatively different species of Centruroides, the 2 morphotypes of C. tecomanus and Centruroides sp. 2, retrieved within the elegans group (Fig. 2). These findings illustrate the complicated patterns of diversification within this buthid genus in Colima. From a taxonomic perspective, if the identity of C. tecomanus is to be clarified, it is now imperative to analyze molecularly exemplars from Tecomán, Colima, which is the type locality of this species.

Thorellius species exhibit a more restricted distribution in Colima. Thorellius intrepidus and T. cristimanus are widely distributed in several states of the Pacific Lowlands (Fig. 6). However, in a recent revision of the genus, 2 species were described, Thorellius wixarika^{González-Santillán and Prendini, 2018} and Thorellius tekuani^{González-Santillán and Prendini, 2018} (^{Fig. 3 of González-Santillán and Prendini, 2018}), that are relevant to these analyses. The former occupies the northwestern territory in Nayarit and Jalisco, whereas the latter inhabits the Balsas Depression of Estado de México, Guerrero, and Michoacán. This observation suggests that the InDRE database contains several misidentifications of both T. intrepidus and T. cristimanus. In fact, the T. wixarika and T. tekuani description was in 2018, and the InDRE database is several years older, hence the complete absence of records. Thus, T. intrepidus inhabits Aguascalientes, Colima, Guanajuato, Jalisco, and Michoacán (Fig. 6A); and T. cristimanus, Colima and Jalisco (^{González-Santillán Prendini, 2018}) (Fig. 6B). From an ecological point of view, we noticed that the InDRE database only has records with elevations below 700 m, implying that these species prefer tropical to subtropical climates in Colima.

^{González-Santillán (2004)} reported Mesomexovis aff. occidentalis as a putative new species for the Biological Station Chamela, Jalisco. Furthermore, ^{González-Santillán and Prendini (2015)} conducted a phylogenetic analysis using morphology and mitochondrial and nuclear DNA, resulting in a topology that suggested that this species was not conspecific to Mesomexovis occidentalis, as ^{Williams (1986)} identified it. ^{López-Granados (2019)} found further morphological evidence to separate this species, but the evidence was never published. The importance of that work is that, for the first time, exemplars of Mesomexovis from Colima that were conspecific with our samples were included in a phylogenetic analysis. Once more, we demonstrate that Mesomexovis sp. 22 is not conspecific with M. occidentalis using molecular evidence, and it requires a formal separation and description (Fig. 2). Unlike Thorellius species in Colima, Mesomexovis is a widespread species inhabiting tropical to montane habitats with a wider range of elevation (Fig. 5B).

The mexicanus group was recently revisited in a monograph with the delimitation of other species groups and the description of 5 new species (^{Contreras-Félix & Francke, 2019}). The paucity of sequences for the mexicanus group in the NCBI database only permitted retrieval of loci for Vaejovis rossmani LP 2027, a species inhabiting the Sierra Madre Oriental in the states of Tamaulipas and Nuevo León and treated in the ^{Contreras-Félix and Francke (2019)} monograph. Our analysis retrieved Vaejovis sp. 32 and Vaejovis rossmani LP 2027 together, which suggests membership of the mexicanus group of Vaejovis. Furthermore, Vaejovis sp. 32 also matches the morphological diagnosis presented in ^{Contreras-Felix and Francke (2019)}. ^{Contreras-Félix et al. (2023)} recorded Vaejovis santibagnezi^{Contreras-Felix and Francke, 2019} in Cerro Grande, where we collected our samples (Fig. 5B). However, the authors failed to present morphological evidence to justify this conclusion. We compared our specimens with the geographically closest V. monticola deposited in the CNAN and with V. santibagnezi and found significant morphological differences. Currently, we are preparing a contribution where we propose the description of a new species. Following this idea, we are inclined to think that, like C. possanii, Vaejovis sp. 32 is also microendemic because of its limitation to disperse throughout tropical valleys from the Cerro Grande massif to other mountain ranges. Finally, we submit that such an evident geographical barrier may apply to several epigean, non-volant arthropods such as the scorpions.

Finally, 2 species absent in our fieldwork are Konetontli ilitchi^{González-Santillán and Prendini 2015} and Vaejovis janssi^{Williams, 1980}. The latter is endemic to the Socorro Island, part of the Revillagigedo Archipelago (^{Williams, 1980}), and K. ilitchi has been found only inside a cave in the vicinity of Coquimatlán, Colima (^{González-Santillán Prendini, 2015}). These 2 features in the distribution and biology of the species hindered the possibility of collecting and studying them. Future fieldwork may illuminate the taxonomy and distribution of these elusive species. Finally, Vaejovis monticola, another high elevation dweller of the mexicanus species group (^{Contreras-Felix & Francke 2019}) was absent in our collecting trips. Although its type locality cited by ^{Contreras-Felix and Francke (2019)} indicates: “Jalisco, northern side of Nevado de Colima” we submit that it may not be on the southern side in Colima, but extensive fieldwork is needed to confirm this hypothesis.