Revision of the morphology, phylogenetic relationships, behaviour and diversity of the Iberian and Italian ant-like Tachydromia Meigen, 1803 (Diptera: Hybotidae)

Phylogenetic inference, based on five molecular markers (COI, 28S, AATS, 12S, PGD), corroborates the synonymy of the flightless genera Pieltainia Arias, 1919 and Ariasella Gil, 1923 with Tachydromia Meigen, 1803. The secondary structure of the 28S rRNA gene is used for the first time in this family to align the multiple sequences. Molecular and morphological data are largely congruent for all known species of flightless Tachydromia. This paper treats ten western Mediterranean species (nine Iberian and one Italian) in detail, including the description of four new species: T. ebejeri Gonçalves, Grootaert & Andrade sp. nov., T. stenoptera Gonçalves, Grootaert & Andrade sp. nov., T. cantabrica Gonçalves, Grootaert & Andrade sp. nov. and T. nigrohirta Gonçalves, Grootaert & Andrade sp. nov. The male of Tachydromia pieltaini (Gil Collado, 1936) and the female of Tachydromia apterygon 1 European Journal of Taxonomy 732: 1–56 ISSN 2118-9773 https://doi.org/10.5852/ejt.2021.732.1213 www.europeanjournaloftaxonomy.eu 2021 · Gonçalves A.R et al. This work is licensed under a Creative Commons Attribution License (CC BY 4.0).

This paper reviews the fl ightless Tachydromia from the western Mediterranean, the Iberian Peninsula and Italy, comprising ten species in total, including four newly described species from Iberia. They are minute ant-like fl ies, ca 2.5 mm, which occupy specifi c microhabitats among the leaf litter of deciduous forests. One of these ant-like Tachydromia, Tachydromia apterygon Plant & Deeming 2006, is endemic to Italy, while the remaining nine species are restricted to the Iberian Peninsula and the Pyrenees.
Apart from Tachydromia apterygon, which has scale-like wings in both sexes and lacks halteres, fi ve of these species were originally placed in the genera Pieltainia Arias, 1919 andAriasella Gil Collado, 1923, but these two genera have since been placed in synonymy with Tachydromia (Shamshev & Grootaert 2018): Tachydromia iberica (Arias, 1919), Tachydromia semiaptera (Gil Collado, 1923), Tachydromia pieltaini (Gil Collado, 1936), Tachydromia pandellei (Séguy, 1941) and the more recently described Tachydromia lusitanica (Grootaert, Shamshev & Andrade, 2009). Despite being morphologically similar to Tachydromia, they were originally assigned to the above-mentioned genera solely based on the complete absence of halteres and on the extremely reduced or absent wings. The former monospecifi c Pieltainia was characterized by aptery in both sexes, while in Ariasella spp. males are stenopterous, i.e., very narrow but long wings (complete loss of fl ight ability) (Figs 20E,D). The males of the species previously placed in Ariasella have tubular, stalk-like wings with a lobed tip, which, at least in T. lusitanica, are waved during mating (Andrade 2011). Females of both T. semiaptera and T. lusitanica are micropterous (Figs 20F,21E), with wings reduced to a minute bilobed scale, while females of T. pandellei and T. pieltaini are stenopterous but lack the developed lobed tip present in males (Fig. 21C).
To our knowledge, there is no published information on the immature stages, but adult Tachydromia are predators of small fl ies, mostly soft-bodied 'nematocera'. They are commonly encountered in large numbers, swiftly running either on vertical or horizontal surfaces such as tree trunks, logs, low vegetation and river margins (Chvála 1970), and often even fully winged species are reluctant to fl y, but instead run to avoid predators. In addition to the Iberian and Italian ant-like species, frequent cursorial habits are also taken to an extreme in only three other previously described species of Tachydromia. These have evolved to become completely fl ightless by lacking functional wings, while still conserving the halteres: Tachydromia rossica Shamshev, 1994 (Russian Far East and Mongolia) has stalk-like wings, without the lobed tip whereas Tachydromia microptera (Loew, 1864) and Tachydromia schnitteri Stark, 1996 (central Europe) are brachypterous (venation and shape similar to macropterous Tachydromia). However, wing reduction is present in at least 20 dipteran families (Hackman 1964) and it is frequent in Hybotidae (Grootaert & Shamshev 2008). This phenomenon may be explained by several factors, including the stability of the habitat and cryptic behaviour (Hackman 1964).
Due to a combination of their small size, cryptic habitat preferences and relatively short period of activity, most Iberian ant-like species of Tachydromia have been poorly studied at the morphological, behavioural, systematic and ecological levels. The original descriptions were generally superfi cial and lacked information on important taxonomic characters, such as the male terminalia. The species distribution was also only known from either the type locality, or just a few localities, generally based on GONÇALVES A.R. et al., Revision of Iberian and Italian ant-like Tachydromia a few specimens, and in the case of T. pieltaini, only the female was known and described. Additionally, the types of T. iberica, T. semiaptera and T. pieltaini seem to be lost and no holotype was assigned to T. pandellei. There is, however, a noteworthy exception to this scenario that is the case of T. lusitanica, whose description includes the male terminalia and other important details of taxonomic relevance. Furthermore, several important observations on its behaviour were published by Andrade (2011). The only other observations were the brief notes by Arias (1919) on Tachydromia iberica, describing its great voracity by preying on fungus gnats (Diptera, Sciaridae Billberg, 1820) and its agility, resembling small ants quickly running on the ground. A close phylogenetic relationship between Ariasella and Tachydromia was recovered by J. Mortelmans (MSc thesis, unpublished data) using several genes (COI, 18S, 28S, EF-1 alpha and PGD). In this study, the single studied species of 'Ariasella', currently T. pieltaini, was resolved embedded among species of Tachydromia. Nagy et al. (2013) resolved T. pieltaini again within the studied species of Tachydromia in a simple Neighbour-Joining tree reconstruction using the 5'-end of the mitochondrial COI gene. Until now, T. iberica (previously regarded as the monotypic genus Pieltainia), has not been included in any phylogenetic inferences.
In the present study we pursue a more inclusive phylogenetic analysis by including all species previously regarded as Ariasella, Tachydromia iberica from three geographically distant localities, and four newly described species. We used fi ve different molecular markers: the entire mitochondrial protein-coding gene of the cytochrome c oxidase subunit I (COI), the partial mitochondrial ribosomal 12S rRNA gene, the partial nuclear protein-coding genes PGD and AATS, and the partial nuclear ribosomal 28S rRNA gene. Furthermore, the present study uses a modern morphological analysis (e.g., by including SEM imaging) in the (re)description of the target taxa, including in the description of the male of T. pieltaini, and a lectotype for T. pandellei is designated. Additionally, we considerably improve the knowledge on the geographic distribution of these species as well as provide new observations on mating behaviour. This research has the overall aims of: i) describing new fl ightless ant-like Iberian species; ii) inferring the phylogenetic relationships among the fl ightless ant-like Iberian species; iii) understanding the phylogenetic relationship between Tachydromia and the species previously placed in genera Pieltainia and Ariasella; and fi nally iv) shedding light on the wing evolution of the ant-like Iberian species.

Sampling, species distribution and specimen deposition
Specimens were manually collected with a vial after being directly searched for on the leaf litter. Afterwards, they were preserved either dry-mounted, in 70% ethanol or absolute ethanol (for molecular analysis). In the cases where type material is lost, species were identifi ed based on the illustrations and on the original descriptions previously published (Arias 1919;Gil 1923Gil , 1936Séguy 1941;Plant & Deeming 2006;Grootaert et al. 2009). In most cases, additional specimens were collected at their type localities or very close by. Distribution data was compiled from published records and based on recent sampling efforts. Several sampling expeditions took place across the Iberian Peninsula and the Pyrenees (Fig. 1) as well as the Apennine Mountains, mostly from January to July (2013)(2014)(2015)(2016)(2017)(2018). Historical and new occurrence records were plotted using QGIS ver. 2.18.16, with the coordinates projected in the reference system Madrid 1870 (Madrid), code 4903. The raster map dataset (Natural Earth II with Shaded Relief, Water, and Drainages) was obtained from https://www.naturalearthdata.com. Specimens not assigned to a formal repository are placed in the fi rst author's collection.

Morphological analysis
Terms used for adult structures follow those of Cumming & Wood (2017). Male terminalia were macerated in hot (around 90°C) 10% KOH for 10 minutes and left at room temperature for several hours, cleared in ethanol and immersed in glycerine for long-term preservation. Drawings of the terminalia were made with a camera lucida attached to a Leitz Laborlux 12 compound microscope. Drawings of the wing were made based on direct observation by using a stereo microscope, SteREO Lumar.V12 (Carl Zeiss MicroImaging). Images of minute wings and spurs were obtained using a scanning electron microscope JSM-5200LV (JEOL, Ltd). Photographs of pinned specimens were taken with a Canon EOS 7D mounted on a P-51 Cam-Lift (Dun Inc.) and exported with the help of Adobe Lightroom (ver. 5.6). Afterwards, the images were stacked using the software Zerene Stacker ver. 1.04. Photographs of live specimens were taken with a Fujifi lm FinePix HS10, with a Raynox DCR-250 macro lens.
The body length was measured from the frons to the end of the terminalia in lateral view. The wing length was measured from the wingtip to its base. Measurements were taken using dried specimens. In the descriptions, the right and left side of the male terminalia are based on the unrotated position viewed posteriorly, such that in the illustrations the right surstylus appears on the reader's left side and vice versa. All male terminalia are fi gured in their unrotated position.   Hackman (1964) proposed a terminology regarding the degree of wing reduction in Diptera. Brachypterous means reduced wing, shorter than abdomen, broad and more or less blunt, not permitting fl ight, with at least radial veins distinct. Stenopterous are specimens with wings very narrow but sometimes long, not permitting fl ight, with at least radial veins distinct. Micropterous are species with wing reduced to small appendage of varying shape with at most traces of the radial vein. And, fi nally, apterous species have wings at most represented by a minute scale, at most carrying some setae or totally absent. Recently, Roháček (2012) added the term submacropterous for species with wings similarly shaped, including venation, as in normal macropterous specimens, but distinctly shorter, darker, about as long as the abdomen. Here, we propose an emendation to this terminology where the term micropterous assumes a more inclusive defi nition: wing reduced to a small appendage, stub or minute scale, of varying shape with at most traces of the radial veins present. Furthermore, we change the term apterous, with its defi nition being restricted to wings totally absent, without any traces of its presence.

Observations on mating behaviour
Observations in captivity were made for Tachydromia semiaptera (population of Spain, Cáceres, Las Villuercas) and Tachydromia iberica (population of Portugal, Leiria, Arrimal). These species were selected because of their level of wing reduction: the former species is stenopterous, with very fl exible stalk-like wings (in opposition to the rigid wings of the other stenopterous Iberian ant-like species); and the latter is apterous. The mating behaviour of a stenopterous species with rigid stalk-like wings, T. lusitanica, was described by Andrade (2011). Specimens were kept in a small terrarium (27 × 15 × 17 cm), with a section of 4 cm deep soil, including leaf litter. In two separated instances, each species (10 ♂♂, 10 ♀♀) was kept and observed for fi ve consecutive days. Observations were made using a Fujifi lm FinePix HS20 EXR, with a Raynox DCR-250 macro lens. Specimens of Sciaridae were offered as a food source. To test intraspecifi c interactions in pre-defi ned conditions (e.g., quantity of specimens, sex), specimens were temporarily taken from the terrarium and put inside transparent plastic vials. Various types of interactions were observed but here, only the mating behaviour is described because of its relevance to help understand the evolutionary importance of wing reduction/modifi cation.
These observations were aimed at obtaining qualitative descriptions of behaviour.

Molecular analysis
Taxon sampling The taxon sampling included 35 specimens belonging to 31 different species: 30 Hybotidae and one Empididae Latreille, 1809. We sequenced and analysed for the fi rst time nine species of Tachydromia, namely eight Iberian ant-like Tachydromia as well as Tachydromia apterygon. In other words, the analysis covered all the nine previously described and newly described species of the Iberian fl ightless ant-like Tachydromia. Additionally, it included four macropterous and one apterous species of Tachydromia assigned to three species-groups sensu Chvála (1970): the terricola, connexa and arrogans groups. The abovementioned taxa constitute the ingroup. Representatives of each of the six subfamilies of Hybotidae were included as outgroups, as were the tribes within Tachydromiinae sensu Sinclair & Cumming (2006). Empis tessellata Fabricius, 1794 (Empididae) was used to root the tree. Table 1 lists the species included in the analysis, as well as the collection data. The samples were made available either by obtaining freshly collected specimens or from the German Barcode of Life (GBOL) DNA collection preserved in the ZFMK BioBank.

Molecular marker selection
Three nuclear genes (28S, PGD, AATS) and two mitochondrial genes (COI and 12S) were selected to be sequenced. The choice of these genes was based on their different evolutionary rates and on the existence of primers (Table 2) used in groups of insects, including some specifi c for Diptera.  DNA extraction and sequencing Entire specimens, ethanol preserved and stored at -20°C, were used for DNA extraction. Extractions were carried out on complete specimens using the Dneasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA) to extract total nucleic acids following the manufacturer's instructions; samples were resuspended in 50-100 μl ultra-pure water instead of in the supplied elution buffer. Entire specimens were preserved and labelled as DNA voucher specimens for the purpose of morphological studies and deposited in ZFMK. The Qiagen Multiplex PCR Kit (Qiagen, Valencia, CA, USA) was used to carry out the PCRs. PCRs (20 μl) included 2.5 μl DNA extract, 1.6 μl of each primer (at 10 pmol/μl), 2 μl of Q-solution, 10 μl of Qiagen Multiplex-Mix and 2.3 μl of ultra-pure water. PCR amplifi cations were carried out with a Biometra PCR Thermal Cycler. PCR 'touchdown' programs are listed in Table 3, primers used for amplifying and sequencing the molecular markers are listed in Table 4. Amplifi ed DNA was visualized on 1.5% agarose gels for inspection of amplifi ed products. PCR products were cleaned up using the commercially available QIAquick PCR Purifi cation Kit (Qiagen, Valencia, CA, USA). Bi-directional Sanger sequencing reactions were carried out by Macrogen Inc. using the same primer pairs as those used for PCR reaction.

Sequence alignment
The sequences were edited for base-calling errors and assembled using Geneious R7 (ver. 7.1.3, Biomatters Ltd). To detect contamination, sequences were compared with known sequences of relatives as well as the GenBank database using nBLAST (NCBI Resource Coordinators 2016). The alignment of the protein-coding COI gene was done manually, and it was not necessary to include gaps in this alignment. The genes 12S, AATS and PDG were aligned by multiple alignment using fast Fourier transform (MAFFT) program (Katoh et al. 2005(Katoh et al. , 2009 ver. 7, which implements iterative refi nement methods (Katoh & Standley 2013). The E-INS-i algorithm was selected for 12S rRNa gene as it is optimized for small-scale alignments and is recommended for sequences with multiple conserved domains and long gaps (Katoh et al. 2009). For the potein-coding genes, the algorithm G-INS-i was used and the resulted alignment, then, checked using the aminoacid translation. The AATS sequences contained a gappy area with some small alignable regions at the beginning of the 5' end. Thus, the fi rst sequenced 206 nucleotides, most likely an intron, were deleted. The alignment of 28S gene was done using its secondary structure, as explained by Kjer (1995) and Kjer et al. (2009). This is the fi rst time that the secondary structure of the 28S rRNA gene is used for the Hybotidae. Two versions of the 28S alignment were produced in order to explore the infl uence of the Regions of expansion and contraction (RECs) and the Regions of slipped-strand compensation (RSCs) in the inferred topology. The REC and RSC regions, which could not be unambiguously aligned, were kept in one version of the alignment (non-trimmed) and deleted in the other version (trimmed).

Sequence analysis
Uncorrected pair-wise evolutionary distance among the Iberian ant-like species (11 nucleotide sequences belonging to nine taxa) was determined by using the p-distance substitution model in the software program MEGA7 (Kumar et al. 2016) (Table 5).

Phylogenetic analyses
For the combined dataset, two different phylogenetic analyses using maximum likelihood (ML) and Bayesian inference (BI) were performed. For both analyses, the molecular data set was divided into 11 partitions: fi rst, second and third codon positions of COI, AATS and PGD, and the entire 28S and 12S genes. The best choice of model was determined for each partition using jModelTest ver. 2. 1.5 (Darriba et al. 2012) under the Akaike Information Criterion (AIC), as recommended by Posada & Buckley (2004), and the data was analysed under the recommended models (see Table 4).
ML analyses were performed using the Genetic Algorithm for Rapid Likelihood Inference, Garli ver. 2.01 (Zwickl 2006). Forty-six independent runs (46 different runs with the command searchreps = 1) were conducted using the scorethreshforterm = 0.05 and signifi canttopochange = 0.001 settings and using the automated stopping criterion, terminating the search when the ln score remained constant for

Sequence characteristics
Aligned nucleotide lengths of gene fragments were of 1504 bp for COI, 518 bp for AATS, 382 bp for 12S, 766 bp for PGD, 838 bp for non-trimmed 28S and 791 bp for trimmed 28S. The total datasets included 4008 nucleotide characters (non-trimmed 28S) and 3961 nucleotide characters (trimmed 28S). Of the 35 specimens included in the analysis, sequences were obtained from 35 individuals, representing the selected 31 taxa, for COI (100% of the data), 19 for 12S (54%), 33 for 28S (94%), 23 for AATS (66%) and 25 for PGD (71%). Species of Tachydromia (i.e., ingroup) were overall successfully sequenced and the sequencing only failed regarding PGD in T. nigrohirta sp. nov. and in T. cantabrica sp. nov., AATS and 28S in T. apterygon. Among the Iberian ant-like Tachydromia, the uncorrected pairwise sequence divergences (p) for the COI gene varied from 8% between T. lusitanica and T. nigrohirta sp. nov., to 12% between T. semiaptera and T. pandellei, and between T. semiaptera and T. pieltaini. The COI p-distance at intraspecifi c (interpopulational) level regarding T. iberica varied from 4% (between Leiria and Madrid) to 5% (between Huelva and Madrid, Huelva and Leiria). The obtained ATTS sequences of species of Platypalpus and Tachydromia (but only those not originally placed in the genera Pieltainia and Ariasella) have an insertion of three nucleotides close to the 3' end of the amplifi ed region. A nexus fi le with the fi nal alignment is provided as Supp. fi le 2 and Supp. fi le 3, and two additional text fi les are also given in Supp. fi le 4 and Supp. fi le 5, with the original structural alignment for 28S.

Reconstruction of phylogenetic relationships
The ML tree, non-trimmed 28S version, with the best likelihood score (ln L = 29397.646621) is presented in Fig. 2. Although the taxonomic sampling was focused on the diversity and phylogenetic relationships within the Iberian fl ightless ant-like Tachydromia and their systematic position to the remaining Tachydromia, other tribes within Tachydromiinae were included in our analyses. Hence, we would like to comment on their inferred relationships. Regarding the ingroup, Tachydromia was recovered paraphyletic with respect to the species originally assigned to former genera Ariasella and Pieltainia. The evolutionary relationships within the ingroup (Tachydromia + Ariasella + Pieltainia) depicted in our results are of a very well-supported grouping with many internal nodes with high to moderate support. The clade (Tachydromia connexa Meigen, 1822 + Tachydromia terricola Zetterstedt, 1819) and representing the connexa and terricola species groups, respectively, was resolved as the sister clade to the group containing the remaining Tachydromia, including the Iberian species. Within these remaining Tachydromia (Tachydromia arrogans (Linnaeus, 1761) + Tachydromia umbrarum Haliday, 1833) representatives of the arrogans and annulimana species group was resolved as a sister clade to the fl ightless species with very high support values (BS = 100; PP = 1). Tachydromia apterygon was placed in the cluster containing all fl ightless species, which is supported by moderate values of bootstrap and high values of posterior probability (BS = 88; PP = 1), indicating that it is more phylogenetically related to the arrogans than to the terricola species group. The Iberian fl ightless ant-like species are a well-supported group with internal nodes generally with high support and with short length branches. Tachydromia semiaptera, originally the type species of the former genus Ariasella, was recovered with a very high support (BS = 97; PP = 1), being the sister taxon to the remaining Iberian species. Moreover, Tachydromia iberica, originally the type species of the monotypic genus Pieltainia, appears as being closer related to the remaining Iberian species, than Tachydromia semiaptera. Tachydromia iberica is represented by specimens obtained from its three main occurrence localities. While there is a signifi cant geographic distance and small differences concerning the male terminalia morphology (see Figs 6-8), these populations are phylogenetically closely related, as can be denoted by the very short branch length. Tachydromia pandellei and T. pieltaini were recovered as sister species with a very high Bayesian posterior probability (PP = 1) but only moderate bootstrap support values (BS = 75). The previously undescribed species were placed in the same grouping (PP = 1; BS = 84), being closely related. Tachydromia nigrohirta sp. nov. and T. lusitanica were recovered as sister species (BS = 80; PP = 0.98). The most related species to the clade (T. nigrohirta sp. nov. + T. lusitanica) are T. cantabrica sp. nov. and T. ebejeri sp. nov.; however, the phylogenetic relationships between these taxa remain unclear due to the lack of statistical support. The phylogenetic position of T. stenoptera sp. nov. was recovered, with a high support (BS = 84; PP = 1), as a sister taxon to [T. cantabrica sp. nov. + (T. ebejeri sp. nov. + (T. nigrohirta sp. nov. + T. lusitanica))].
The topology of the majority-rule consensus tree resulting from Bayesian inference agrees with the most likely tree regarding all taxa of Tachydromiinae. Additionally, and regarding the datasets with trimmed and non-trimmed 28S, both analyses (by ML and BI) inferred the same topology for the Tachydromiinae ( Bootstrap support values (below) and Bayesian posterior probabilities (above) are depicted at the nodes (only > 50 or > 0.5, respectively). Abbreviations: BS = Bootstrap support values; PP = Bayesian posterior probabilities. A greyscale is used to highlight the ingroup, where the darkest shade of grey highlights the Iberian fl ightless ant-like species of Tachydromia Meigen, 1803, followed by a lighter shade which includes T. apterygon Plant & Deeming, 2006, hence representing all the fl ightless species occurring in southern Europe and, fi nally, the lighter shade covers all Tachydromia analysed, including the macropterous species assigned to different species groups sensu Chvála (1970). The white bar indicates the species originally assigned to genus Pieltainia Arias, 1919, while the grey bars indicate the taxa originally assigned to genus different species-groups sensu Chvála (1970

General diagnosis of the Iberian and Italian ant-like species of Tachydromia
The Iberian and Italian ant-like Tachydromia species can be jointly characterized by lacking halteres and functional wings (apterous, micropterous or stenopterous). Considering their overall morphology (e.g., shape and colouration of palpus and pubescence of legs), it can be said that Iberian species are generally more related to the arrogans species group than to any other; however, various species differ from this group in having a shorter stylus 1.5 times as lo ng as the antenna (T. semiaptera, T. pieltaini and T. pandellei) and/or for having male genitalia large and globular (T. stenoptera sp. nov.). Plant & Deeming (2006) suggested that T. apterygon is more related to terricola group. They are quick and agile runners and often move their abdomens up and downwards, resembling ants or certain fl ightless parasitic wasps.
Thorax black with portion of episternum densely covered with grey microtrichia. Legs with colour pattern of yellowish and dark brown to black; fore tibia with ciliation of short coarse setae. Mid tibia with anteroventral spur at apex, produced into short apical elongation, about as long as diameter of tibia at apex (Fig. 19J); spur with several distinct setae and, on distal half margin, a few short, stout, curved spinose setae. Abdomen black, covered with grey microtrichia and short setae, longer at hind margin of last sternite.

Description
Female (previously undescribed) Similar to male. Micropterous, wing round, brownish, bilobate, apical margin of posterior lobe bearing 1 long seta, anterior lobe bearing 3 shorter setae, lobes covered with grey microtrichia. Differs from male in following: legs mainly covered with microtrichia and regular rows of setae; spur on mid tibia absent; cercus pale brown with grey microtrichia and long setae; a few longer setae on apical sternites.

Distribution and habitat
Italy. Currently, this species is only known from the central Apennine Mountains (Fig. 23). It occurs on the edge of forests of Fagus sylvatica L. (Fig. 24), on the leaf litter and short herbaceous vegetation, where it can be abundant. It has also been found in mixed deciduous woodland at lower altitudes, inside the forest, on the leaf litter.

Remarks
Despite the sampling effort at the end of July, the species was not found at the mixed deciduous woodland type locality. However, it was still very abundant at higher altitudes. The type specimen was collected at the beginning of July, so by the end of the month, it may already be too hot and dry for this species to survive at lower altitudes.

Diagnosis
Overall dark. Wing dimorphic: male stenopterous; wing with lobed distal apex, no veins distinguishable, dark brown for most part, with black and white lobe; female micropterous, wing bilobate, with 1 seta on each lobe. Palpi, proboscis and antennae black; postpedicel sub-conical, ca 1.5 times as long as pedicel. Legs with a colour pattern of yellowish and dark brown to black; male fore tibia with ciliation of long hairlike setae. Abdomen black, tergites and sternites with evenly distributed setae, covered with grey microtrichia. It shares similarities with T. nigrohirta sp. nov. and T. stenoptera sp. nov., but can be distinguished from these species by the darker leg colouration, lobed distal apex of male wing without any trace of apical digitation, sub-conical postpedicel, and male terminalia.

Etymology
This species is named after the Spanish Cantabrian mountain range, where it was found. HEAD. Face and frons largely glabrous, shiny black; ocellar tubercle shiny anteriorly and fi nely pollinose posteriorly. 1 pair lateroclinate ocellars, same length as postpedicel, no posterior ocellars. 1 pair proclinate verticals, ca ¼ shorter than length of postpedicel, and numerous setae on occiput; no postocular ciliae present; mouth margin covered by grey microtrichia and with long setulae.

Material examined
ANTENNA. Black, covered by grey microtrichia. Scape as long as wide; pedicel twice as long as wide, with apical circlet of black setae, ventrals longer; postpedicel sub-conical, ca 1.5 times as long as pedicel, with numerous pale setulae longer dorsally and apically; stylus almost twice as long as scape, pedicel and postpedicel combined.
PALPUS. Elongate oval, about 3 times as long as wide. Black in ground colour with grey microtrichia, dorsally clothed in strong, long silvery-white and black setae, bearing one strong black sub-apical seta slightly shorter than palpus; less numerous setae present ventrally with distinct seta on ventral apex.
LEGS. Fore femur stout and infl ated basally, mid femur less so and hind femur least. Black with the following pale brown/yellowish: basal portion of coxae, trochanters, basal portion of mid and hind femora, distal apical surface of fore and mid femora, knees, basal 4/5 surface of tarsomere 1 and basal ⅔ surface of tarsomere 2. Legs mostly covered with regular rows of numerous black and pale setulae, except for coxae, trochanters and posteroventral surface of femora; anteroventral surface of coxae, femora (except most of ventral surface) and tibiae covered by grey microtrichia; coxae with several short, downward directed, dorsal anteroapical setae. Femora with a few distinct apical setae. Fore tibia bearing sparse ciliation of hair-like setae (shorter than maximum diameter of tibia) with slightly curved tips, and row of long, erect, setae present at its distal ⅓ posterodorsal surface. Tarsomere 1 of fore tarsus with strong and long (twice tarsomere 1 width) setae; at posterior distal surface with row of distinct, short, black setae. Mid tibia with spur at anteroventral apex, with barely noticeable apical elongation, about as long as diameter of tibia at apex (Fig. 19H); spur has several distinct setae and, on distal half margin, fi ve short, stout, curved, spinose setae.
ABDOMEN. Black, uniformly covered with grey microtrichia. Tergites and sternites with evenly distributed setae; apical sternites with long, distinct setae. TERMINALIA (Fig. 4). Subglobular, mostly covered with grey microtrichia, dark brown. Right surstylus short with anterior margin produced into short slender projection, mostly glabrous, without microtrichia, bearing stronger setae mostly on dorsal margin, with several (ca 8) distinct denticles, most of which aggregated on distal apex. Right epandrial lamella with irregularly placed long setae, with patch of similar setae on lateral surface. Cerci of similar length, both with long setae of unequal length, enclosed in epandrial lamellae. Left epandrial lamella slightly shorter than left cercus, bilobed, with a few long setulae on apical margin. Left surstylus subrectangular with rounded apical margin, 3 times as long as wide, distinctively longer than cercus, densely covered with grey microtrichia with equally long setae on apical margin.

Female
Similar to male, except for following features: micropterous, with wings round, brownish, bilobate, bearing 1 seta on apical margin of each lobe, covered by grey microtrichia (Fig. 20D); legs covered with regular rows of setae; spur on mid tibia absent; cercus pale brown with grey microtrichia and setae; no longer setae on apical sternites.

Distribution
Spain. Currently only known from the Cantabrian Mountains.
ABDOMEN. First tergite with differentiated, strong posteromarginal setae. TERMINALIA (Fig. 5). Right surstylus covered with long setulae, without microtrichia, bearing stronger setae mostly on dorsal margin, with several (ca 10) distinct denticles, ⅔ of which aggregated on distal apex. Right cerci with broad tip, left narrowing distally, both with long setae of unequal length on apical half, enclosed in epandrial lamellae. Left epandrial lamella ⅔ as long as left cercus. Left surstylus 2 times as long as wide, longer than cerci, with short setae of equal length on apical margin.

Female
Similar to male, including being micropterous (Fig. 20B), except for following features: no distinctive hairs or setae on fore femur; spur on mid tibia absent; cercus pale brown with grey microtrichia and setae; no longer setae on apical sternites.

Distribution
Portugal and Spain. Mostly distributed in the northwestern Iberia (North and Central Portugal and Galicia), with just one locality south of the Tagus river (Santa Maria de Marvão). (Arias, 1919)

Diagnosis
Predominantly dark and glabrous. Wings, or any traces of it, absent. Palpi, proboscis and antennae black. Postpedicel subconical, 2 times longer than pedicel; stylus 2 times as long as scape, pedicel and postpedicel combined. Legs with dark brown to black and yellow colour pattern. Abdomen black, tergites and sternites mostly covered by grey microtrichia and by sparse, very short, setae; apical sternites with long, distinct, setae.
ANTENNA. Postpedicel 2 times as long as pedicel.
PALPUS. Less numerous setae present ventrally with a few distinct setae.

WING. Absent.
LEGS. Dark brown to black with following yellow: trochanters, apical half of fore and mid femora, dorsal surface and basal ⅓ of fore and mid tibiae, most of tarsomeres 1 and 2, except for black apical margin. Ciliation of hair-like setae absent from fore tibia. Spur absent. TERMINALIA (Figs 6-8). Right surstylus long, mostly glabrous, except for stronger setae on dorsal and distal margins. Right epandrial lamella with long, irregularly placed setae on apical ⅓. Cerci with long setae of unequal length mostly placed on apical ⅓, Left epandrial lamella ⅓ shorter than left cercus, deeply bilobed. Left surstylus almost twice as long as wide, ⅓ shorter than cerci.

Female
Except for terminalia, no distinguishing features between sexes. Cerci pale brown with grey microtrichia and setae.

Variability
Main sources of variability are related to leg colouration, number of prescutellar setae and shape of right surstylus. In all populations, it is possible to fi nd specimens with dark tarsi, instead of yellow. Regarding the right surstylus, in the southern population, this structure is usually long, narrowing distally (Fig. 6); in the Spanish central system, specimens usually have a very short, sub-triangular, right surstylus (Fig. 7); the Portuguese specimens, often have a very long and narrow surstylus (Fig. 8). However, it is possible to fi nd variability concerning the shapes of right surstylus within the same population. Concerning the number of prescutellar setae, the usual number is 2 setae, but in all populations analysed it is possible to observe specimens with several prescutellar setae (ca 6).

Distribution
Spain and Portugal. It is known from the Serranía de Cuenca in eastern Iberia, Guadarrama mountains in the centre, in the foothills of the Sierra Morena mountains in the south, and in the Portuguese provinces of Beira Litoral and Estremadura in the west. Tachydromia lusitanica Shamshev & Grootaert 2018: 425 (comb. nov.).

Diagnosis
Small (2.25-2.40 mm (n = 3)), slender, dark and hirsute species. Wing dimorphic: male stenopterous; bilobed distal apex, with very long digitation on distal margin (Fig. 20E). Stalk-like process dark brown on basal half, pale on distal half; lobed apex mostly black, digitation hyaline with black distal ⅓ margin. Female micropterous, wing bilobate, squamiform, apical margin of each lobe bearing setae -the posterior lobe has 2 long setae and the anterior one 2 short setae (Fig. 20F). Wing veins not distinguishable in either sex. Occiput largely covered by grey microtrichia; 1 pair of long lateroclinate ocellars; 1 pair of proclinate verticals; palpi, proboscis and antennae black; postpedicel lanceolate, 2.5 times as long as pedicel; stylus 2 times as long as scape, pedicel and postpedicel combined. Thorax black with episternum densely covered with grey microtrichia, appearing sub-triangular in lateral view. Postpronotal lobe bearing 7 setae laterally, ca 9 acrostichal setae, irregular biseriate, of equal length; notopleuron with 1 strong, prominent seta plus 5 small setae; 2 strong prescutellar setae present laterally; scutellum with 2 similar setae medially. Legs mainly dark with pattern of yellow and brown to black; male fore tibia with dense ciliation of long hair-like setae; male mid with anteroventral apical spur produced into a barely noticeable apical elongation, about as long as diameter of tibia at apex (Fig. 19A); spur has several distinct setae and, on the distal half margin, 4 short, stout, curved, spinose setae. Abdomen black, tergites and sternites overall uniformly covered by grey microtrichia and long setae; apical sternites with longer setae. Terminalia ( Fig. 9) subglobular, mostly covered with grey microtrichia, dark brown. Right surstylus produced into short projection, bearing long setae at the lateral margins and ca 11 denticles. Right cerci slightly longer than left, both enclosed in epandrial lamellae. Left epandrial lamella bilobed.   ANTENNA. Postpedicel lanceolate, 2 times as long as pedicel. Stylus 1.5 times as long as scape, pedicel and postpedicel combined.
WING. Stalk-like portion dark brown on basal half, pale on distal half (Fig. 20G).

GONÇALVES A.R. et al., Revision of Iberian and Italian ant-like Tachydromia
LEGS. Black, except for yellowish to brown trochanters, knees and tarsomeres 1 and 2. Fore tibia with dense ciliation of long hair-like setae (distinctively larger than maximum diameter of tibia) with curved tips. Mid tibia anteroventral apical spur with distinct apical elongation (Fig. 19I), on basal half margin with 4 short, stout, curved spinose setae. TERMINALIA (Fig. 10). Left epandrial lamella ⅓ shorter than left cercus, deeply bilobed. Left surstylus 2 times as long as wide.

Female
Similar to male, except for following features: micropterous, wing round, brownish, bilobate, apical margin of each lobe bearing setae, posterior lobe with 1 long seta and the anterior with 3 short setae, both covered with grey microtrichia (Fig. 20H); legs covered with regular rows of setae; spur on mid tibia absent; cercus pale brown with grey microtrichia and setae; no longer setae on apical sternites.

Diagnosis
Overall shiny black with distinct yellow parts. Wing dimorphic: male stenopterous; wing with cordiform (French populations) or lobed apex (Spanish populations); stalk-like portion yellow to light brown; lobed European Journal of Taxonomy 732: 1-56 (2021) apex with basal half hyaline, distal half black; female stenopterous, stalk-like portion yellowish with minute lobed apex pale brown. Antennae yellow with brownish postpedicel; palpi yellow; proboscis black; postpedicel roundish, approximately as long as pedicel. Legs overall yellow with brownish hind femora and tibiae, tarsomeres yellow and black. Abdomen black, tergites and sternites overall covered with grey microtrichia and setae.

Redescription
Similar to Tachydromia cantabrica sp. nov. except for the following:
ANTENNA. Scape and pedicel yellow, postpedicel brown. Postpedicel oval, 2 times as long as wide and slightly longer than pedicel; stylus slightly more than 3 times as long as scape, pedicel and postpedicel combined.
WING. Lobed apex cordiform. Stalk-like portion rigid, yellow except for brown basal portion; lobed apex with basal half hyaline (appearing silver under certain light conditions), distal half black (Fig. 21A).

GONÇALVES A.R. et al., Revision of Iberian and Italian ant-like Tachydromia
LEGS. Overall yellow, except for following brown to black portions: posterodorsal surface of coxae, apical ¼ dorsal surface, posterior and anterior surfaces of hind femur, most of hind tibia, tarsomere 5 of fore and mid-legs, distal apex of tarsomeres 1 and 2, distal ⅓ of tarsomere 3, distal ¾ of tarsomeres 4 and 5. Mid tibia anteroventral apical spur produced into distinct and long apical elongation (Fig. 19B), longer than diameter of tibia at apex; mostly glabrous ventrally but dorsally covered by microtrichia, with a few distinct setae and, on basal half margin, 3 short, stout, curved setae. Elongation mostly glabrous, with a few setae arising marginally and apically. TERMINALIA (Fig. 11). Right epandrial lamella with long, erect setae mostly on apical half. Right surstylus bifurcated into short projection and very long and slender projection; mostly glabrous but with long, curved lateral setae and 2 denticle-like setae at distal margin of longer projection; short projection with similar setae at distal margin. Right cercus slightly longer than left, both enclosed in epandrial lamellae. Left epandrial lamella as long as left cercus. Left surstylus subcylindrical, narrowing proximately, 4 times as long as wide, as long as right cercus, with setae on lateral margins and apical ⅓.

Female
Similar to male, except for following features: stenopterous (length of wing (n = 2): 0.7-0.8 mm), with very minute lobed apex, stalk-like portion yellowish with lobe light brown, covered by grey microtrichia (Fig. 21C); legs covered with regular rows of setae, with distinct circle of setae on fore femur and anteroapical setae on coxae, but no other distinctive hairs or setae; spur on mid tibia absent; cercus pale brown with grey microtrichia and setae.

Variability
There is a variability in the shape of the wing lobe in males between the populations sampled in the French Pyrenees compared with those from Spain. In the former, the lobed apex is distinctively large and cordiform, while in the latter the lobed apex is comparatively smaller and oval. Fig. 11. Terminalia of the topotype of Tachydromia pandellei (Séguy, 1941)  Tachydromia pieltaini (Gil Collado, 1936) Figs 1, 12, 16E-F, 18G-H, 19C, 21B-C Ariasella pieltaini Gil Collado, 1936: 191.

Diagnosis
Shiny black with distinct yellow parts. Wing dimorphic: male stenopterous; wing with oval apex; stalklike process yellow to light brown; lobed apex with hyaline cells, veins and distal margin black; female stenopterous, stalk-like portion yellowish with minute lobed apex pale brown. Proboscis and antennae black; palpi yellow; postpedicel roundish, approximately as long as pedicel. Legs overall yellow only with the distal tarsomeres black. Abdomen black, tergites and sternites overall covered with setae; grey microtrichia present at last apical tergite and most sternites, resulting in shiny appearance of dorsal surface.
ANTENNA. Yellow. Postpedicel roundish, as long as wide and, approximately as long as pedicel; stylus almost 4 times as long as scape, pedicel and postpedicel combined.
PALPUS. About 4 times as long as wide. Yellow in ground colour.

GONÇALVES A.R. et al., Revision of Iberian and Italian ant-like Tachydromia
LEGS. Overall yellow, with the following brown to black: dorsal strip on fore and mid femora, posterior coxae, tarsomere 5 of fore and mid-legs, apical portions of tarsomeres 2 and 3, tarsomeres 4 and 5 of hind leg. Mid trochanter with several black anteroventral marginal setae. Mid femur with small roundish posteroventral projection on basal surface with several black denticle-like setae and ca 3 black setae. Mid tibia anteroventral apical spur produced into distinct, long elongation (Fig. 19C); slightly longer than diameter of tibia at apex. Spur covered with microtrichia and > 10 short, blunt denticle-like setae; additionally, with 6 longer, stout, curved setae along lateral margin. Elongation mostly glabrous with a few setae arising marginally and apically.
ABDOMEN. Shiny black, with grey microtrichia covering only last tergite and almost all sternites, except for sternites 1 and 2. Tergites and sternites overall covered with setae, except for dorsal surface of all tergites and most of sternite 1 (distal margin with a few setae); apical sternite and tergite with longer setae. TERMINALIA (Fig. 12). Right surstylus short with irregular margins, narrowing distally, bearing long setae, with several (ca 8) distinct denticle-like setae aggregated on distal margin. Right cercus longer than left. Right cercus with setae on apical ⅓; left cercus with setae of unequal lengths on apical half and 2 long apical setae. Left epandrial lamella ⅓ shorter than left cercus, with setae of similar length on apical half of ventral surface and on apex of dorsal margin. Left surstylus longer than right cercus, with very long setae on lateral and apical margins.

Diagnosis
Overall dark species with distinct yellow parts. Wing dimorphic: male stenopterous; oblong apex; stalklike portion light brown with distinct black spot on lobed apex; female micropterous, with wing round, brownish. Palpi, proboscis and antennae black; postpedicel roundish, approximately as long as pedicel.
Legs with colour pattern of yellowish and dark brown to black. Male legs with following features: fore tibia black, stout and infl ated; mid tibia bearing two projections and row of fi ve spinose setae on posteroventral basal surface. Abdomen black, overall covered with microtrichia and sparse setae; two apical sternites with long setae. HEAD. Lateroclinate ocellars 2.5 times as long as postpedicel. Proclinate verticals 2 times as long as postpedicel.
ANTENNA. Light brown. Postpedicel roundish, as long as wide; stylus more than 4 times as long as scape, pedicel and postpedicel combined.
LEGS. Fore tibia stout and infl ated distally. Overall dark brown to black, following portions yellow: trochanters, anterior apical half and apex of fore and mid tibiae, dorsal strip and apical half of fore and mid tibiae, fore and mid tarsomeres 1, 2, 3, basal ⅔ 4, hind tarsomeres 1 and 2. Fore tibia with rows of long, erect setae; mid tibia bearing row of 5 strong, spine-like posteroventral basal setae. Mid tibia with 2 anteroventral apical projections (Fig. 19E-F): 1 slightly shorter than diameter of tibia at apex; clothed in microtrichia, several distinct setae and 2 short, stout, curved spinose setae on margin of distal half; second projection as long as diameter of tibia at apex with a few long setae. European Journal of Taxonomy 732: 1-56 (2021) TERMINALIA (Fig. 13). Right epandrial lamella with long, erect setae on apical half. Right surstylus short, narrowing distally, with sparse microtrichia, bearing long, irregularly curved setae, with several (ca 6) distinct denticle-like setae, aggregated on distal margin. Right cercus with setae on apical half and distinct apical seta; left cercus with apical setae of unequal lengths. Right cercus longer than left, both enclosed in epandrial lamellae. Left epandrial lamella shorter than left cercus, deeply bilobed. Left surstylus 1.5 times as long as wide, approximately as long as right cercus, with very long setae, lateroclinate, facing inwards.

Female
Similar to male, except for following features: micropterous, with wing round, brownish, bearing 1 long seta and 1 short seta on apical margin, covered by grey microtrichia (Fig. 21E); legs covered with regular rows of setae, with distinct apical setae on fore femur and anteroapical setae on coxae; without other distinctive setae; spurs on mid tibia absent; cercus pale brown with grey microtrichia and setae.

Distribution
Spain and Portugal. It is known from several localities along the Iberian Central System, and from the Montes de Toledo mountain range, the northeast of Portugal and from one locality in the Spanish province of Zamora near the border with Portugal. Gonçalves, Grootaert & Andrade sp. nov. urn:lsid:zoobank.org:act:AB1CBE5D-22D9-4153-9EE5-FA711157B977 Figs 1, 14, 15G-H, 17G-H, 19G, 21F-G

Diagnosis
Overall very dark. Wing dimorphic: male stenopterous; wing with slightly lobed distal apex, dark for most part and lobe black and white; female micropterous, wing bilobate. Palpi, proboscis and antennae black; postpedicel lanceolate, ca 1.5 times as long as pedicel. Legs almost completely black. Abdomen black, tergites mostly glabrous and without grey microtrichia.

Etymology
The name of this species means 'narrow wing' and derives from the combination of two Greek words: the prefi x steno-(stenos), meaning 'narrow', with the suffi x -ptera (pterá), meaning 'wing'. Hence, the name refl ects the very narrow lobed distal apex of the male wing.

Description
Similar to Tachydromia cantabrica sp. nov. except for the following:
LEGS. Overall black; ventral and dorsal posterior surface of coxae, trochanter, basal 0.1 ventral surface of femora and knees, paler. Fore tibia bearing sparse ciliation of hair-like setae slightly longer than maximum diameter of tibia. Mid tibia anteroventral apical spur with 4 short, stout, curved spinose setae on distal half margin.

Female
Similar to male, except for following features: micropterous, with wings round, brownish wings, bilobate, bearing two setae on apical margin of each lobe, covered by grey microtrichia (Fig. 21G); legs covered with regular rows of setae, with distinct apical setae on fore femur, without other distinctive setae; apical projection on mid tibia absent; cercus pale brown with grey microtrichia and setae; without longer setae on apical sternites.

Distribution
Portugal. Only known from the western end of the Iberian Central System. It has not yet been found in Spain, but given that it was found very close to the border, it is very likely that it will eventually be found in this country.

Notes on habitat, distribution, and phenology
The fi eldwork yielded a signifi cant amount of new data regarding both the geographic distribution of most species (Fig. 1) and their (micro)habitat. Currently, the following number of localities are known for each species: T. cantabrica sp. nov. is known from 3 localities, T. ebejeri sp. nov. from 23, T. lusitanica from 41, T. nigrohirta sp. nov. from 3, T. pandellei from 13, T. pieltaini from 5, T. semiaptera from 27 and T. stenoptera sp. nov. from 7. These species occur in temperate and sub-mediterranean regions, mainly in mountains, where they often inhabit forests dominated by deciduous or marcescent trees, such as Quercus robur L., Q. pyrenaica Willd. and Fagus sylvatica L. They are only to be found among the European Journal of Taxonomy 732: 1-56 (2021) leaf litter and very low herbaceous vegetation. Furthermore, most Iberian ant-like Tachydromia tend to become active in early spring (mostly from February until May); however, species as T. pandellei and T. pieltaini only occur at higher altitude and, hence, become active much later in the year (likely May onwards). These species often occur in the forests of Fagus sylvatica but seem to be present only on the edge of the forest. The Iberian ant-like Tachydromia are often found among other fl ightless insects, especially ants, such as Lasius Fabricius, 1804. However, no direct predation on Tachydromia by ants was ever observed. The Tachydromia themselves seem to prey only on adult Sciaridae (fungus gnats) and Chironomidae (non-biting midges).

Observations on mating behaviour of T. semiaptera and T. iberica
Successful mating attempts have only been observed in both species when the male was in possession of a prey. The movements performed by a male interested in copulating, as described below, are the same even if it does not hold a prey, with the additional fact that he chases the female. Usually, the female attempts to steal the prey and either the male allows it and immediately mounts her or, showing no interest in copulating, fl ees away with the prey. In the last scenario, the male may, at most, briefl y move its stretched abdomen vertically. In the case of T. iberica he also may move the mid legs sideways, as observed when he is confronted with another male or any perceived danger.
When the female approaches the male to steal the prey, the male interested in copulating stretches his abdomen and waves it up vertically but, additionally, moves his forelegs sideways. The male offers no resistance when the female takes the prey and, instead, readily mounts her back, beginning the mating attempt. Instantly, the male places his forelegs above the female's eye level, while the mid legs are placed in the female's waist in an attempt to grab it. At the beginning of the copula, the male moves his forelegs sideways at the female's eye level, while, with his posterior legs, rubs his own last tergite and twitches it. The movement of the forelegs is very similar to the movement of the wings described in T. lusitanica by Andrade (2011). When their genitalia seem to be locked, the male keeps himself very still, always with his forelegs above the female's eye level. The female then may move around, sometimes very fast, while keeping the prey with the male on top. Frequently, the male of T. iberica lets himself be carried around by the female in a tail-to-tail position. When the specimens disconnect, both individuals usually fi ght vigorously for the prey.

Phylogenetic relationships and systematics
The phylogenetic relationships within Tachydromiinae have been studied previously, but mostly at the tribe and generic level (e.g., Nagy et al. 2013) with the specifi c diversity within the genus yet to be studied in further detail. This is the fi rst time that all the Iberian and Italian ant-like species have been sequenced, and their phylogenetic relationship as well as the evolutionary relationships between every ant-like species and the genus Tachydromia analysed.
The inferred phylogenetic position of the type species of Ariasella as sister taxon to the remaining Ariasella species and Pieltainia supports the conclusion by Shamshev & Grootaert (2018) that the absence of functional wings has no phylogenetic value at generic level. The placement of the two genera within the clade of Tachydromia corroborates their synonymy under Tachydromia. The Iberian ant-like species of Tachydromia form an evolutionary closely related group and the inferred phylogenetic relationships are in accordance with the morphological characters. Hence, T. semiaptera is resolved as the sister taxon to all other Iberian species as its peculiar morphology suggests. This taxon possesses a combination of characters in the male (i.e., two projections on mid tibia, lack of ciliation of long hairlike setae on the fore tibia, roundish postpedicel and fl exible wing) that are not found in any other species. A spur-like structure (Fig. 19) on the mid tibia of the male is present in all Iberian and Italian species, apart from T. iberica, which completely lacks it. As described in the morphological analysis, this structure is, for most species, composed of a basal section bearing denticles and a side projection. In the case of T. semiaptera, the basal section and projection are totally separated or, at least, their separation is much deeper than in other species. To our knowledge, this is the only case in the genus Tachydromia where this deep or total separation occurs. Furthermore, this structure likely has a evolutionary relevance, as the males of the species in which mating was observed, grab the females by using the mid legs.
Tachydromia iberica is a species that lacks many of the characters with taxonomic relevance found in other species as it is apterous in both sexes, does not have spur-like structures or characteristic setae (e.g., ciliation of long hair-like setae on the fore tibia). Some differences at the terminalia, mainly concerning the right surstylus, were found among three geographically distant populations. However, these differences are of degree rather than regarding the basic form and have less taxonomic relevance. Additionally, the specimens sequenced are still very much related as recovered in the phylogenetic analysis, where they are separated by short branches.
Furthermore, and even though we conducted a total evidence phylogenetic analysis, it seems relevant to comment on the COI p-divergences as this gene is often used to distinguish species, despite some recognized issues (e.g., Meier et al. 2006). While the genetic diversity at species level within Hybotidae remains scarcely studied, large intraspecifi c COI p-divergences are not uncommon in Hybotidae (see Nagy et al. 2013). In the mentioned study, for example, an intraspecifi c divergence of up to 5.48% was found within the studied Tachydromia. In the specifi c case of T. iberica, the high divergence found at interpopulational level (4-5%) may be representative of vicariance processes, especially as no signifi cant morphological or ecological differences were found to substantiate the separation of the populations as distinct species.
Concerning the sister-species T. pieltaini and T. pandellei, these two species are morphologically similar but with very clear differences that allow for a straightforward species separation. In this case, the two more relevant characters are perhaps the wing pattern and shape as well as the terminalia (in males), which are very different. These morphological differences are highly congruent with the molecular data, as these species are separated by relatively long branches in the inferred phylogenetic tree.
The four newly described species all belong to the same cluster where T. lusitanica was resolved. With the exception of the micropterous T. ebejeri sp. nov., the male wing, in this species cluster, is stenopterous with his lobed tip divided into a white/transparent basal portion and black distally, but it is in T. lusitanica where it presents a more remarkable shape by bearing a distal projection. Tachydromia nigrohirta sp. nov. was inferred as a sister taxon to T. lusitanica and this pair has less morphological differences in comparison with, for example, T. pieltaini and T. pandellei. There is not enough phylogenetic signal in the studied molecular markers to fully resolve the relationships between T. stenoptera sp. nov., T. cantabrica sp. nov. and T. ebejeri sp. nov., which may be a result from a rapid speciation process.
Regarding the placement of the Iberian ant-like species within the species groups previously defi ned for Tachydromia, it is possible to associate them more closely with the arrogans species group, although this association is uncertain due to the lack of developed wings and other morphological characteristics which do not fi t into the group defi nition, such as the shorter stylus and/or the large and globular male genitalia. The overall morphology -despite the lack of developed wings -places the Italian Tachydromia apterygon within the terricola species group, as suggested by Plant & Deeming (2006). However, the molecular data clearly places it as being more closely related to the arrogans species group. Therefore, further molecular studies are necessary to understand the phylogenetic relationships of T. apterygon and the monophyly of the suggested Tachydromia species groups based on morphological data.
Concerning the slight differences of the support values between the trimmed and nontrimmed datasets, we assume that both RECs and RSCs have some effect on the branch support values, not on the topology as primarily thought. In our results, the inferred topology using the trimmed version of the 28S alignment receives, overall, slightly lower support values than the non-trimmed, indicating that the ambiguous regions of the 28S gene provides some support for the best tree.
A hypothesis on the evolution of these species can be proposed based on the recovered phylogenetic relationships and knowledge about the behaviour and habitat. While the divergence events are not dated, the short internodes in the cluster of the Iberian Tachydromia may very well refl ect a rapid succession of splitting processes. It is possible that geographically separated populations adapted differently to similar ecological conditions, hence simultaneously evolving different sexual signalling strategies. Thus, differences may have arisen following vicariance processes and now these species show distinct behaviours and life histories, which may block the gene fl ow between them in sympatry.

Wing reduction/modifi cation in Tachydromia and its evolutionary signifi cance
As mentioned before, the lack of functional wings is a common phenomenon in Diptera and several hypotheses for fl ightlessness in insects have been discussed, including habitat persistence and cryptic behaviours (Hackman 1964). Roff (1990) concluded that the loss of wings is common in spatially and temporally homogeneous environments and that there is considerable evidence for an increase in fl ightlessness with altitude. The Iberian ant-like species are mostly present on mountains and in deciduous forests, which are known to be stable habitats (Roff 1990), their microhabitat being restricted to the leaf litter. Overall, the evolution of fl ightlessness in Tachydromia is currently known from 13 species (including the Iberian and Italian taxa), out of some 100 described species. It is expected that more cases may be discovered, especially if fi eld work is carried out at high altitude in temperate cold environments. Regarding the evolution of the wings in the Iberian ant-like Tachydromia, we hypothesize that two processes may have taken place: one is the wing reduction itself, which may happen in both sexes, and another is the wing modifi cation with a sexual signalling function, which only happens in the male. Therefore, in some species only wing reduction to different degrees occurred, from the stenopterous species to the apterous, with the micropterous inbetween. In certain cases, the female is stenopterous without a distinct lobed tip on its wing, while the male wing has a very distinct and large lobed tip (see: Tachydromia pieltaini, Tachydromia pandellei, Fig. 21A-C). However, more commonly, a more distinct sexual dimorphism is present, where the female is micropterous and the male is stenopterous.
It is interesting that the sexual signalling process does not happen the same way in all species, as observed in the mating behaviour of T. semiaptera. Contrary to T. lusitanica, the male of this species does not wave its wings during copula, nor do the wings seem to have any other function. Moreover, its wings are much less rigid than the ones of all the remaining related species, so it is possible that the stiffness of the stalk-like portion of the wing is an important part of the wing modifi cation process, making them easier to wave. Hence, while the wings of the stenopterous species may look similar at fi rst glance, they actually differ in their morphology and might have different functions.
In T. semiaptera only a process of wing reduction seems to be taking place. Nonetheless, this species has a very distinct feature, which has the same signalling function as that of the wing in T. lusitanica. Its fore tibia -which is black, stout and infl ated -is waved in front of the female's eyes during copula. A similar signalling method is used by T. iberica (waving the forelegs) but this species, instead of a modifi ed tibia, simply has a yellow and black foreleg. Furthermore, both species lack the ciliation of hair-like setae on the male foreleg, present in T. lusitanica, whose forelegs were observed to intertwine with those of the female (Andrade 2011). This behaviour was not observed in T. semiaptera or T. iberica.
Taking all information into account, at least in the case of T. lusitanica, wing reduction seems to be coupled with an important modifi cation of sexual relevance, and it appears that both ecological and sexual pressures are driving its evolution. Similar situations are likely to be present in the other species that show a distinct sexual dimorphism in the wing morphology, but this remains to be tested.

Remarks on the habitat and distribution with implications on conservation
As previously mentioned, the ant-like species of Tachydromia occur in regions of temperate and submediterranean bioclimatic infl uence. In these bioclimates, the potential natural vegetation is composed of broad-leaved deciduous or marcescent species and, often, the forests are dominated by oak trees, such as Quercus robur L. and Quercus pyrenaica Willd (Fig. 25) (Costa et al. 2005). At higher altitude, the oak trees are substituted by Fagus sylvatica L., a deciduous tree that forms monospecifi c stands with a closed canopy, which blocks the penetration of sunlight and creates very shaded environments (Costa et al. 2005). The soil of these deciduous and marcescent forests tends to have a deep layer of leaf litter. This layer of leaf litter, with some degree of light penetration, high humidity and rich in organic matter, GONÇALVES A.R. et al., Revision of Iberian and Italian ant-like Tachydromia is the typical microhabitat where ant-like Tachydromia can be found. With the exception of the species occurring at high altitude, the adults of Iberian ant-like Tachydromia become active at the beginning of the spring, when the trees are still without leaves, and the sunlight can directly reach the ground. The fl ies can then be found both inside the forest and along the forest edge.
Tachydromia pandellei and T. pieltaini often occur in habitats where the forest is often dominated by Fagus sylvatica, and they probably only live along the edge of the forest because it provides the presence of leaf litter combined with higher light penetration than microhabitats under dense canopy. The leaf litter itself is a humid, stable environment and likely offers shelter from predators and the weather elements (e.g., wind, rain). It is also the habitat for prey, as these species mainly feed on saprophilous fl ies that are dependent on organic matter.
In general, at a macroclimatic scale, the ant-like species of Tachydromia seem to have relatively similar ecological requirements and some species do co-occur in the same habitats (Fig. 2). This is, for example, the case of T. semiaptera, which co-occurs with T. ebejeri sp. nov., T. nigrohirta sp. nov. and T. iberica. However, T. semiaptera seems to become active later in the year than T. ebejeri sp. nov. and T. nigrohirta sp. nov., and does not overlap at a temporal scale with T. ebejeri sp. nov. in the same habitats. We suspect that further fi eldwork likely will increase the geographic ranges reported for most species, especially of T. semiaptera, T. cantabrica sp. nov., T. pandellei and T. pieltaini. However, given the previous sampling efforts, T. stenoptera sp. nov. and T. nigrohirta sp. nov. appear to be genuinely restricted to a smaller area of occurrence.
Most populations of the Iberian ant-like Tachydromia exist in highly fragmented landscapes given their occurrence in temperate and humid forested habitats. This is especially true for the species with populations in the southern half of the Iberian Peninsula, where they are circumscribed to small submediterranean bioclimatic islands surrounded by a dry and hot environment. This is the case for the southernmost populations of T. iberica, T. semiaptera and T. ebejeri sp. nov., but even the populations in the northern regions occur in fragmented habitats. The habitat fragmentation is often caused by disturbances of anthropogenic origin, such as the artifi cialization of the land, agricultural intensifi cation, substitution of the natural vegetation by monocultures and invasive species, as well as induced fi res. The climate is also expected to become warmer, greatly reducing the area occupied by temperate and submediterranean trees, as Querqus robur and Q. pyrenaica (Benito Garzón et al. 2008). Climate change, coupled with habitat fragmentation, will almost certainly imperil the conservation of the Iberian ant-like species of Tachydromia.

Supplementary material:
Supp. fi le 1: Fig.1. Maximum-likelihood tree based on the combined dataset (COI, trimmed 28S, 12S, AATS and PGD) using Garli ver. 2.01.1067 and the structural alignment for 28S. Bootstrap support values are depicted at the nodes (only > 50 or > 0.5, respectively). BS = Bootstrap support values. A greyscale is used to highlight the ingroup, where the darkest shade of grey highlights the Iberian fl ightless ant-like Tachydromia species, followed by a lighter shade which includes T. apterygon, hence representing all the fl ightless species occurring in southern Europe and, fi nally, the lighter shade covers all Tachydromia analysed, including the macropterous species assigned to different species-groups sensu Chvála (1970). Fig. 2. Bayesian Inference tree based on the combined dataset (COI, trimmed 28S, 12S, AATS and PGD) using MrBayes ver. 3.2.6 and the structural alignment for 28S. Bayesian posterior probabilities are depicted at the nodes (only > 50 or > 0.5, respectively). PP = Bayesian posterior probabilities. A greyscale is used to highlight the ingroup, where the darkest shade of grey highlights the Iberian fl ightless ant-like Tachydromia species, followed by a lighter shade which includes T. apterygon, hence representing all the fl ightess species occurring in southern Europe and, fi nally, the lighter shade covers all Tachydromia analysed, including the macropterous species assigned to different species-groups sensu Chvála (1970).