The phylogenetic placement of the enigmatic pseudoscorpion family Menthidae (Pseudoscorpiones): a revised superfamily assignment based on new molecular data

Introduction

The arachnid order Pseudoscorpiones is an ancient, moderately diverse group of arachnids with ~4060 species and 479 genera (World Pseudoscorpiones Catalog 2025). The smallest adults are less than 1 mm in body length and the largest are ~1 cm. Although their fossil record is incomplete, the earliest fossils are cuticular fragments from the Middle Devonian (Schawaller et al. 1991; Judson 2012), with all subsequent fossils from the Mesozoic and Cenozoic being readily assigned to modern families (Harms and Dunlop 2017), suggesting deep origins for all clades (see also Benavides et al. 2019). Judson (2012) regarded Dracochela Schawaller, Shear & Bonamo, 1991 as a stem-group pseudoscorpion, but Benavides et al. (2019) included it in its own suborder of crown-group pseudoscorpions. Pseudoscorpions have a worldwide distribution and can be found in most terrestrial ecosystems.

The past three decades has seen various phylogenetic hypotheses for the entire order with a morphology-based cladistic treatment (Harvey 1992), and molecular phylogenies based on three genes (Murienne et al. 2008; Arabi et al. 2012; Harvey et al. 2016b) or transcriptomes (Benavides et al. 2019). Despite iterative modifications to the classification, the results of these analyses are now converging on a phylogenetic classification that comprises three modern suborders: Heterosphyronida with Chthoniidae and Pseudotyrannochthoniidae (superfamily Chthonioidea); Atoposphyronida with Feaellidae and Pseudogarypidae (Feaelloidea); and Iocheirata with six superfamilies including Neobisioidea (Bochicidae, Gymnobisiidae, Hyidae, Ideoroncidae, Neobisiidae, Parahyidae and Syarinidae), Garypoidea (Garypidae, Geogarypidae, Hesperolpiidae, Menthidae and Olpiidae), Garypinoidea (Garypinidae and Larcidae), Sternophoroidea (Sternophoridae), Cheiridioidea (Cheiridiidae and Pseudochiridiidae) and Cheliferoidea (Atemnidae, Cheliferidae, Chernetidae and Withiidae). Confidence in the strength of the various phylogenies has grown over time such that changes to the pseudoscorpion classification have been implemented to better reflect the phylogenetic treatments. For example, Murienne et al. (2008) recovered a paraphyletic Garypoidea, which was only formalised based on the more comprehensive phylogenomic treatment by Benavides et al. (2019) with the recognition of a separate superfamily Garypinoidea in addition to a reduced Garypoidea.

Previous molecular studies have included species from most pseudoscorpion families, thus providing relatively good coverage across the phylogenetic breadth of the order. However, two enigmatic families – Menthidae and Pseudochiridiidae – have so far lacked any molecular data and have never been included in molecular analyses. Both are rarely encountered and are sparsely distributed around the world.

Ever since the genus Menthus Chamberlin, 1930 and the family Menthidae were first erected (Chamberlin 1930), they have been mostly considered to be garypoids with the greatest morphological similarity (Fig. 1) to Olpiidae (Chamberlin 1931; Beier 1932; Muchmore 1982; Harvey and Muchmore 1990). The only exception was the inclusion of Olpiidae and Menthidae in their own superfamily Olpioidea based on the results of a morphological cladistic analysis (Harvey 1992). Olpioidea was abandoned by Benavides et al. (2019), and both families were returned to the Garypoidea. Menthids differ from the other garypoid families in a suite of unusual features such as the presence of 11 trichobothria on the fixed chelal finger and hand, the absence of the venom gland in the movable chelal finger and the presence of a specialised articulation joint between coxae II and III (Chamberlin 1930, 1931; Harvey and Muchmore 1990; Harvey 1992). Nevertheless, their affinities have never been questioned.

Fig. 1.

Specimen of Thenmus aigialites Harvey, 1990, dorsal.

The recent collection of fresh specimens of the menthid Thenmus aigialites Harvey, 1990 from eastern Australia has allowed us to test the phylogenetic position of this enigmatic family, and to compare it with independent transcriptome data obtained from Menthus californicus Chamberlin, 1930 from south-western USA. To our surprise, menthids were recovered within the superfamily Neobisioidea, which prompted a renewed examination of their morphological traits.

Methods

We analysed two independent data sets, incorporating a wide variety of different pseudoscorpion taxa, with the aim of robustly assessing the phylogenetic position of Menthidae. These included a Sanger sequencing-based phylogeny with new sequence data and a novel menthid transcriptome that was added to the phylogenomic analysis of Benavides et al. (2019). The results from these two molecular independent data sets (different taxa, different markers) were used to investigate the systematic position of Menthidae. Chitrella or any of its putative relatives (i.e. members of the syarinid subfamily Chitrellinae) could not be included in the phylogenomic analysis.

The first dataset was an extension of previous Sanger amplicon studies using the mitochondrial cytochrome c oxidase subunit 1 (COI), and two nuclear genes, the small ribosomal subunit (18S rRNA, 18S) and the large ribosomal subunit (28S rRNA, 28S) (Murienne et al. 2008; Arabi et al. 2012; Harvey et al. 2015, 2016a, 2016b, 2020), with the inclusion of two menthids. Thenmus aigialites was sequenced for all three markers, based on a newly collected specimen from eastern Queensland. Menthus californicus (MCZ:IZ:149277) was also included based on a 1047-bp fragment of 18S (GenBank Accession: MW463272.1) extracted from the transcriptome reads. As the monophyly of Iocheirata has been convincingly demonstrated in all previous Sanger-based (Murienne et al. 2008; Arabi et al. 2012; Harvey et al. 2016b) and transcriptomic (Benavides et al. 2019; Ontano et al. 2021; Ballesteros et al. 2022) studies, only two outgroups from the other suborders were added, Austrochthonius sp. representing Heterosphyronida and Neopseudogarypus scutellatus Morris, 1948 for Atoposphyronida. Menthidae is unequivocally a member of the Iocheirata due to the presence of a venom apparatus in the fixed chelal finger (Chamberlin 1930; Harvey 1992). The number of exemplars in Cheliferoidea was also reduced (six exemplars; four families), as there was no evidence that menthids are closely related to them based on morphological criteria, including the presence of two pairs of eyes in most menthids, separate metatarsi and tarsi and the lack of spermathecae in the female genital system (Chamberlin 1931; Harvey 1992). The number of exemplars in other superfamilies were as follows: Neobisioidea (28 exemplars; 6 families); Garypoidea (27 exemplars; 4 families); Garypinoidea (10 exemplars; 2 families); Sternophoroidea (1 exemplar; 1 family); and Cheiridioidea (1 exemplar; 1 family) (Table 1).

For the Sanger-sequencing dataset, fragments of COI, 18S and 28S were sequenced using standard amplicon methodologies, as outlined in previous studies on pseudoscorpions (Murienne et al. 2008; Harvey et al. 2015, 2016a, 2016b, 2020) – with the exception mentioned above for Menthus californicus. The molecular methods used for the polymerase chain reaction (PCR) amplification of each of these genes for more recently sequenced specimens followed Harvey et al. (2015, 2016a, 2020), with PCR purification and Sanger bidirectional sequencing conducted by the Australian Genome Research Facility (AGRF; Perth). Chromatograms were edited using the Geneious software package (ver. 2021.2, Biomatters Ltd, see https://www.geneious.com/, accessed June 2021), and resulting nucleotide sequences for all taxa are deposited in GenBank (Table 1). The sequences were aligned using the MAFFT (ver. 7.490, see https://mafft.cbrc.jp/alignment/software/; Katoh et al. 2002; Katoh and Standley 2013) plug-in within Geneious with the default settings. A total of 3657 aligned base pairs (bp) were analysed, including 658 bp for COI, 1795 bp for 18S and 1204 bp for 28S. The concatenated alignment was analysed using maximum likelihood (ML) and Bayesian inference (BI). ML analysis was run in the web version of IQ-TREE (ver. 2.2.0, see https://github.com/iqtree/iqtree2; Trifinopoulos et al. 2016; Minh et al. 2020), and the data were partitioned per gene. The substitution model option was set to Auto, and branch support was assessed with 5000 replicates of the ultrafast bootstrap approximation (Hoang et al. 2018) (bootstrap support or BS hereafter). BI was implemented in MrBayes (ver. 3.2.7, see https://github.com/NBISweden/MrBayes/; Ronquist et al. 2012); each partition had a unique GTR substitution model with gamma-distributed rate variation across sites and a proportion of invariant sites (GTR + Γ + Ι). The analysis consisted of two runs of 5 million generations each, with four chains, sampling every 100 generations, evaluating convergence of the independent runs through the standard deviation of split frequencies in Tracer (ver. 1.7.2, see https://github.com/beast-dev/tracer/releases/tag/v1.7.2; Rambaut et al. 2018). A consensus tree was estimated after discarding a burn-in of 10% of the sampled trees.

Both Sanger-sequencing analyses were rooted on Austrochthonius sp., a member of the superfamily Chthonioidea (World Pseudoscorpiones Catalog 2025), which represents the sister taxon to the suborders Atoposphyronida + Iocheirata (Benavides et al. 2019).

The phylotranscriptomic dataset was an expanded version of the transcriptomic study undertaken by Benavides et al. (2019), with the addition of Menthus californicus (see https://mczbase.mcz.harvard.edu/guid/MCZ:IZ:149277), and the exclusion of Cybella gelanggi Harvey, 2018 due to the low number of BUSCO genes present in this assembly (Supplementary Table S1). The majority of the 26 pseudoscorpion families were included in the study, with 23 included in both the phylotranscriptomic and Sanger datasets. The exceptions were Bochicidae for which we lacked Sanger data, and Amblyolpiidae and Olpiidae for which we lacked transcriptomic data.

The raw reads of M. californicus were processed following the same pipeline as for the other taxa in Benavides et al. (2019). BUSCO (Manni et al. 2021a) was used to quantify the completeness of our assemblies (Table S1) by comparing them with the Arthropoda database (Manni et al. 2021b).

The Orthologous Matrix (OMA) database standalone (ver. 2.6.0, see https://github.com/DessimozLab/OmaStandalone; Altenhoff et al. 2019) was used to infer orthologous genes across our samples; each OMA-generated orthogroup was aligned individually using MAFFT (ver. 7.309; Katoh and Standley 2013) eliminating positions with more than 80% missing data with a custom script (trimEnds.sh). One matrix was assembled targeting a minimum gene occupancy of 75% meaning that an OMA orthogroup was selected if present in at least 75% of the taxa. Our resulting matrix consisted of 235 loci (58,041 aa).

The transcriptomic matrix was analysed under maximum likelihood (ML) and Bayesian inference (BI). ML analysis was run in IQ-TREE multicore version (ver. 2.2.2.7, see https://github.com/iqtree/iqtree2) on the FAS Research Computer cluster, performing a gene-partitioned analysis (PART) with the option MFP + MERGE, which selects the best-fit partitioning scheme by merging partitions to reduce over-parameterisation and increase model fit, and including the LG4 mixture models. Nodal support was calculated using 1000 ultrafast bootstrap replicates (-bb 1000; Hoang et al. 2018). The unpartitioned matrix was used to run a BI analysis in ExaBayes (ver. 1.5.1, see https://github.com/aberer/exabayes; Aberer et al. 2014) under the default GTR model (Lartillot and Philippe 2004). The Markov Chain Monte Carlo (MCMC) was configured with two runs and two coupled chains per independent run, each with one cold and three hot chains sampling every 500 until the average standard deviation of split frequencies (ASDSF) was <0.02. Tracer (ver. 1.7.2; Rambaut et al. 2018) was used to check parameter convergence and effective sample size (ESS) of the MCMC runs. The runs converged after ~1.7 × 10⁶ generations and the first 25% of trees were discarded as burn-in.

Results

For the discussion, we draw on the Sanger and transcriptome IQ-TREE analyses as our reference trees, but the resulting topologies for all our analyses (ML – Sanger, Fig. 2; BI – Sanger, Fig. S2; ML – Transcriptomes, Fig. 3; BI – Transcriptomes, Fig. S3 of the Supplementary material), are virtually identical and unambiguously place Menthidae within the superfamily Neobisioidea, and not within Garypoidea or Olpioidea as previously postulated (Chamberlin 1930, 1931; Harvey 1992; Benavides et al. 2019). In the Sanger-based analysis, Thenmus and Menthus were recovered as sister taxa (Fig. 2) in a clade with Gymnobisiidae and the syarinid Chitrella Beier, 1932 (BS = 93%), which was in turn sister group to all nine genera of Neobisiidae. The phylotranscriptomic analysis (Fig. 3) placed Menthus as a sister group to a clade that includes Gymnobisium (Gymnobisiidae) and Neobisiidae, albeit with 55% BS. The strong concordance between both analyses strongly supports menthids being a neobisioid and a member of the clade that also contains Gymnobisiidae and Neobisiidae.

Fig. 2.

Maximum likelihood phylogeny of selected pseudoscorpions, based on alignment of concatenated cytochrome c oxidase subunit I (COI), 18S rRNA (18S) and 28S rRNA (28S). Bootstrap support values are presented for nodes greater than 80%.

Fig. 3.

Pseudoscorpion phylogeny inferred from the maximum likelihood analysis of 235 lociorthogroups (58,041 aa; 75% occupancy) in IQ-TREE partitioned by gene and model search including LG4 mixture models and accounting for heterotachy (lnL = −1,232,140.595). Nodes without values have maximal support.

The dating analysis infers that menthids diverged from Gymnobisiidae + Neobisiidae in the early Mesozoic with a median expected divergence in the early Jurassic (Fig. 4). Gymnobisiidae and Neobisiidae then diverged in the mid-Mesozoic, and all such dates are consistent with the dating analysis presented by Benavides et al. (2019). Indeed, all pseudoscorpion families have arisen during the Mesozoic, as evidenced by our analysis that is predicated on numerous Cretaceous fossils that are readily assigned to modern families (Table 2).

Fig. 4.

MCMCtree chronogram showing divergence time estimates for Pseudoscorpiones evolution obtained from the analysis of 235 loci (58,041 aa; 75% occupancy) with 95% highest posterior density (HPD) bars under a uniform prior. Numbers on nodes indicate the placement of the fossils used to calibrate the dating analysis.

The superfamily Neobisioidea sensu Benavides et al. (2019) was not found to be monophyletic with various taxa either as sister group to the entire Iocheirata, i.e. Parahyidae and some Syarinidae, or as sister group to Panctenata, i.e. Hyidae, Ideoroncidae and some Syarinidae (Fig. 2).

Garypinidae was rendered paraphyletic by the inclusion of Larcidae and Amblyolpiidae. It should be noted that this clade lacked support in the analysis (Fig. 2), and that a different result was obtained by Gao et al. (2025) using a larger dataset. Similar problems were found by Harvey et al. (2020) whereby the garypid subfamilies Garypinae and Synsphyroninae were non-monophyletic despite very convincing morphological synapomorphies to the contrary.

Discussion

Menthids are one of the rarest pseudoscorpion families with only 5 genera and 12 valid species described to date. Their worldwide distribution is sporadic, with each genus restricted to isolated regions in both the northern and southern hemispheres: Menthus Chamberlin, 1930 (four species) in the south-western USA and Mexico; Oligomenthus Beier, 1962 in Argentina and Chile (two species); Paramenthus Beier, 1963 (two species) and Pseudomenthus Mahnert, 2007 (two species) in the Middle East, Iran and Yemen; and Thenmus Muchmore, 1990 (two species) in northern Australia (Fig. 5). They display a potentially vicariant pattern that is difficult to reconcile with dispersal hypotheses, as menthids are small, seemingly delicate arthropods that rely on humid microclimates that have never been found in a phoretic relationship with other arthropods, unlike some other pseudoscorpion taxa (e.g. Beier 1948; Muchmore 1971). Indeed, menthids are small, fragile pseudoscorpions that do not survive outside of their mesic environment, e.g. after being transferred to glass vials (M. S. Harvey, T. L. Miller and M. G. Rix, unpubl. data). A more plausible explanation is that they are Mesozoic relicts with Pangean origins but have only survived in isolated regions of the world due to subsequent extinction events, a result that remains to be tested by including multiple menthids in a dated phylogeny. However, the split from its sister taxon largely overlaps with a time of a unified Pangea, when the family probably initiated its diversification. This Pangean-wide pattern is also evident in other pseudoscorpion clades such as Pseudotyrannochthoniidae (Harms et al. 2024), Pseudogarypidae (Harvey and Šťáhlavský 2010), Syarinidae (Harvey 1998) and Garypinidae (Harvey and Šťáhlavský 2010). A similar pattern also occurs in some early lineages of the Opiliones superfamily Triaenonychoidea, the family Buemarinoidae (Derkarabetian et al. 2021). The superfamily Neobisioidea is thought to have arisen in the late Paleozoic with all of the families established by the Triassic (Benavides et al. 2019).

Fig. 5.

Map depicting the known records of menthids.

Menthids (Fig. 6) possess some very unusual morphological features (Chamberlin 1931; Muchmore 1982; Harvey 1992). They are the only adult pseudoscorpions that have 11 trichobothria on the fixed chelal finger and hand (Harvey and Muchmore 1990; Harvey 2006; Mahnert 2007) (Fig. 11). Most other pseudoscorpions have eight trichobothria, although all species of Ideoroncidae have supernumerary trichobothria on both chelal fingers (e.g. Mahnert 1984; Harvey and Muchmore 2013; Harvey and Du Preez 2014; Harvey 2016), even in Cretaceous taxa (Geißler et al. 2022). Several families have reduced numbers of trichobothria, most likely due to neoteny (e.g. Harvey 1992), including Neobisiidae, Syarinidae, Garypidae, Geogarypidae, Olpiidae, Garypinidae, Larcidae, Cheiridiidae, Sternophoridae, Chernetidae and Cheliferidae. The condylar articulation between coxae II and III of the prosoma in Menthidae (Chamberlin 1931, fig. 7B, H) (Fig. 9) is unparalleled in other pseudoscorpions, even though its function is unknown.

Fig. 6–11.

Images of Menthus californicus Chamberlin, 1930 (WAM T109850) showing important taxonomic features of menthids: 6, body, dorsal; 7, body, ventral; 8, left movable cheliceral finger, dorsal; 9, cephalothorax, ventral; 10, left pedipalp, dorsal; 11, left chela showing supernumerary trichobothria a, c and d. Scale lines: 0.5 mm.

Despite having been placed in Garypoidea (Chamberlin 1930, 1931), our results solidly support its placement in Neobisioidea. This placement actually makes sense under the light of several morphological characters that menthids share with the rest of the neobisoideans that were not previously given sufficient attention. These include the following six characteritics.

There are several other features that are consistent with a menthid placement in Neobisioidea.

The venom apparatus of menthids is only present in the fixed finger and is completely absent in the movable finger (Fig. 11). In all garypoids and garypinoids, the venom apparatus is present in both fingers, but is often absent in neobisioids (Chamberlin 1931; Harvey 1992). It is lost in the fixed finger of Gymnobisiidae, some Hyidae (the genus Indohya Beier, 1974), some Bochicidae (Vachonium Chamberlin, 1947 and the subfamily Leucohyinae), and from the movable finger in all Neobisiidae, Parahyidae and Syarinidae. The loss of the venom apparatus in the movable finger of menthids is therefore in accord with its inferred relationship as sister to the clade Gymnobisiidae + Neobisiidae + Parahyidae + Syarinidae, with an interesting change of fingers in gymnobisiids. Like Neobisiidae, Parahyidae and Syarinidae, the venom duct of menthids is very short and terminates in the nodus ramosus almost immediately, whereas it is longer in the other families.

The presence of three auxiliary trichobothria on the fixed chelal finger and hand resulting in a total of 11 trichobothria (Fig. 10, 11) is a definitive autapomorphy for adult Menthidae, which is unparalleled in any other pseudoscorpions (Harvey 1992). Most pseudoscorpions have 8 trichobothria although some have reduced numbers due to neoteny, and all Ideoroncidae have supernumerary trichobothria resulting in 17–31 on the fixed chelal finger and hand, and 9–14 on the movable finger (e.g. Mahnert 1984; Harvey and Muchmore 2013; Harvey and Du Preez 2014; Harvey 2016).

The position of trichobothrium ib on the medio-dorsal surface of the hand of menthids resembles that found in some other neobisioids including Ideoroncidae, Bochicidae and most Syarinidae (e.g. Arcanobisiinae, Chitrellinae, and most Ideobisiinae, but not Syarininae) (e.g. Beier 1952; Mahnert 1984; Muchmore 1996, 1998; Zaragoza 2010; Harvey and Muchmore 2013; Gardini 2015; Harvey 2016). In some other neobisioids such as Hyidae, ib is located on the dorsal surface but in a more distal position close to the base of the fingers (Harvey 1993; Harvey et al. 2023). There are no garypoids or garypinoids that have ib situated dorsally on the hand.

Summary phylogeny

Taking into account all of the molecular evidence, it is possible to add Menthidae to a summary phylogeny as sister group to a clade of Gymnobisiidae + Neobisiidae (Fig. 12). The discovery of a group of neobisioid pseudoscorpions with distinctive morphological traits (Chamberlin 1930, 1931; Harvey and Muchmore 1990) and an isolated phylogenetic placement (Fig. 2, 3) suggests that there is much more to discover among the many disparate clades of Neobisioidea. In particular, the non-monophyly of the family Syarinidae remains a vexatious problem that should be best resolved with a more robust phylogenomic dataset and better taxon representation.

Fig. 12.

Summary tree depicting the relationships of pseudoscorpion families based on Benavides et al. (2019) and this study. Suborders Palaeosphyronida, Heterosphyronida, Atoposphyronida and Iocheirata in bold; group Homosphyronida and infraorders Hemictenata and Panctenata in regular font.

Supplementary material

Supplementary material is available online.