Phylogeny of the ant genus Aphaenogaster ( Hymenoptera : Formicidae ) in the Iberian Peninsula , with the description of a new species

A phylogenetic tree of the Iberian Aphaenogaster species - except for A. splendida (Roger) - and a key to the worker caste of all Iberian Aphaenogaster species are proposed. The position of A. striativentris Forel and A. cardenai Espadaler is discussed, stating the possibility that this second species may belong to a new, undescribed genus. Aphaenogaster ulibeli n. sp. is described from the Iberian Peninsula. Its closest relatives are A. gibbosa (Latreille) and A. striativentris. Its habitat seems to be restricted to caducifolia forests in the Western Central Massif. 


Introduction
In 2007 the AIM (Asociación Ibérica de Mirmecología: http://www.mirmiberica.org/)developed an extensive pitfall sampling in 18 sites in the Iberian Peninsula.The coordination and identification of the samples was made by X. Espadaler and K. Gómez.
Two undescribed Aphaenogaster males were collected in a pitfall in the Sierra de Béjar (Salamanca) site by Alberto Sánchez.The males on this genus are strikingly different from one species to another, and allow identification to species level in a relatively easy way.
Reidentification of the Aphaenogaster workers from the site showed that samples previously identified as A. gibbosa were a new species indeed.The finding of new nests

Terminology and Measurements
Specimens were examined and/or measured under a MZ 16A stereo microscope.All images were edited in Lenovo Photo Master and OpenOffice Impress, including those used from AntWeb (2016) with permission.Some specimens examined have alphanumeric codes associated with them (i.e., CASENT#, KG#, XE#) which uniquely identify the specimens for databasing purposes.
PCR products were purified by ammonium acetateethanol precipitation and reconstituted in 10 µl of LTE buffer (10mM Tris, 0,1mM EDTA).Direct Sanger sequencing of amplified fragments was done on both DNA strands using PCR primers.Sequencing was conducted using the Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) following manufacturer's instructions, and samples were loaded onto an ABI 3730XL automated sequencer.

Data analysis
Chromatograms were revised, PCR primers trimmed sequences corresponding to each individual ant assembled and final fasta sequences generated for each analysed ant using the Staden package v1.6.0 (Staden et al., 1998).).Multiple alignment of COI sequences were done using the ClustalW program included in the MEGA 6 software (Tamura et al., 2013).The sequence of Myrmica rugiventris (accession number GQ255171) was also included in the alignment to be used as outgroup in the phylogenetic analysis.

Phylogenetic analysis
MEGA 6 software was used for the phylogenetic analysis.The "Find Best DNA Model" option available in MEGA 6 was used to find the evolutionary model that best fit the data.A phylogenetic tree was constructed using the Maximum Likelihood method based on the General Time Reversible model [1].A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.6888)).The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 46.7811% sites).Node support was evaluated by the Bootstrap method (Felsenstein, 1985) using 500 pseudo replicates of the original data.Bootstrap values were included next to the branches when higher than 50%.

DNA extraction, PCR amplification and sequencing
Total DNA was extracted in each sample separately from three individual ants that had been preserved in 96% ethanol since the collection date.
DNA was extracted following the HotSHOT (Hot Sodium HidrOxide and Tris) method (Truett et al., 2000) using 60 µl of both alkaline lysis and neutralizing reagents.A 710 bp fragment of the 5' region of the mitochondrial gene coding the cytochrome c oxidase subunit 1 (COI) was amplified using primers LCO1490 and HCO2198 described by Folmer et al. (1994).For the PCR reactions, 0.5 µl of the extracted DNA were used in a total reaction volume of 50 µl.Each PCR reaction also contained one unit of Taq polymerase (VWR), 1X buffer, 0.2 mM of each dNTP and 0.2 µM of

Worker
Holotype and Paratypes: CL 1.  HEAD: Antennae 12 segmented with 4 segmented antennal club, all segments longer than wide.Scape cylindrical, longitudinally striated, long-clearly surpassing the occipital border when laid back.Abundant greyish white semi-erect setae present in scape and funiculus, its length similar to scape maximum width.These setae decumbent in the basal zone and gradually rising to be semierect in the apical third.
Head clearly longer than wide with subparallel sides.Eyes located medially, well developed, much wider than scape width and with 14-20 ommatidia in its longest axis.Mandibles triangular, longitudinally striated.Dentition with 4-5 more developed teeth decreasing in size from the apical, and 4-7 smaller denticles following to the basal line.Clypeus emarginated medially, longitudinally striated.In some individuals medial portion of the clypeus between the frontal lobes are smooth and shiny.
MESOSOMA: Promesonotal line in profile view continuous.Mesosoma with clearly demarcated mesopropodeal suture.Spines smaller than the space between them, triangular, directed upwards.Metasternal process variable, a rounded lobe to sharply pointed.
COLORATION: Color blackish brown to black, except mandibles and tarsi, yellowish brown.Some individuals dark brown (possibly recently hatched) SCULPTURATION: Overall background sculpture feebly reticulated, absent on the gaster, which is smooth and shiny.Head with subparallel striae overlapping this pattern, with some transverse smaller striae present, but without creating a reticulum.These striae becoming scarcer on the posterior fourth, where only some feeble striae reach the occipital line and the background reticulation is clearly seen.Lateral sides of head and gula with some scattered striae, but most of the surface only reticulated.Pronotal dorsum mainly smooth with some faint striae present.Pleurae longitudinally striated, that continues on the propodeum as parallel transverse striae.Petiole, postpetiole and gaster smooth and shiny, without striae of any kind.
SETATION: Long, greyish white erect to semi-erect setae abundant overall including gaster tergites and lateral METASOMA: Petiole and postpetiole as in worker, slightly more peaked and some feeble striae maybe present in the posterior face of the petiole.
SCULPTURATION: Lateral sides of the head and gula striated.Dorsal surface of scutum faintly striated with irregular subconcentrical striae, lateral sides smooth, overall appearance smooth and shiny.Scutellum smooth and shiny with mesoscutellum feebly striated.Anepisternum and katepisternum smooth in its anterior half and longitudinally striated in its posterior half.Pronotum transversely striated.
Mandibles feebly striated longitudinally, margin smooth and shiny, armed with four sharp teeth, the apical long and curved, and decreasing in size to the basal tooth.Clypeus emarginated, divided in two sections, the central upper section raised.Eight long, grey hairs located basally in this upper section and covering the mandibles.Clypeal sculpture weakly reticulated, but without rugulae of any kind.Frontal ridges developed, but frontal lobes very small, so that antennal insertions clearly exposed.Eyes with microscopic hairs.
MESOSOMA: Mesoscutum swollen, overhanging the pronotum in dorsal view.Propodeum declivity an almost straight line (approximately 30 degrees with the horizontal) in one specimen, and with a short, vertical slightly convex face before meeting the scutellum in the other.Propodeal lobes very from reduced to non existent.Metasternal process formed by a small, blunt triangle oriented backwards.Between the second coxae and this metasternal process, another two lateral sharply pointed processes similar in size and shape.
COLORATION: Color brown to light brown, except mandibles, antennae and legs, light to yellowish brown.
SCULPTURATION: Head sculpture reticulated with a few isolated striae radiating from the ocelli.Some striae (2-4) between the lobes.Another 2-4 striae running on the frontal lobes upwards, some reaching feebly the lower ocellus.Scutoscutellar, oblique mesopleural sulcus and metapleuropropodeal suture with a transverse rugulated pattern.The rest of the body smooth and shiny.
SETATION: Grey to white long setae present on head, mandibles, dorsal surfaces of mesosoma, petiole, postpetiole and gaster.Absent on the genae, lateral surfaces of the mesosoma and very reduced to non existent on the propodeum.

Phylogenetic position
Phylogenetic analysis.As expected, a 710 bp DNA fragment containing a portion of the mitochondrial COI gene was amplified through PCR from the ants analysed.After removing sequences corresponding to primers used in  COI sequences were obtained for ants from nine Aphaenogaster and one Messor species.Sequences for the different individual ants were identical for any sample, so that a single sequence was finally assigned to each colony.Three different colonies were analyzed for Aphaenogaster ulibeli and Aphaengaster subterranea.The rest of the species are represented by one single colony.The sequences were deposited deposited in GenBank (accession numbers on Table 1).
A phylogenetic tree for the ant genus Aphaenogaster in the Iberian Peninsula was obtained and is summarized in Figure 6.

Discussion
Our analysis support the existence of four different clades into the Aphaenogaster species present in the Iberian Peninsula.
Clade 2 contains the classical Aphaenogaster s. str.species (A.senilis, A. iberica and the mediterranean A. spinosa).This clade is coherent with the A. testaceopilosa species group defined in De Boer (2013)  A. spinosa, from continental Italy, is also included in the analysis.M. rugiventris is used as outgroup.The tree is drawn to scale, with branch lengths measured in the number of substitutions per site site (bar corresponds to 0.02 substitutions per site).Bootstrap values are indicated on branches when higher than 50.Three of the four clades described in the text are indicated.
is therefore a mix of species of belonging to the pallida and subterranea groups as defined in De Boer (2013).
A third, interesting clade 1 is coherent with the A. gibbosa species group (De Boer, 2013) and includes the new species with A. striativentris and A. gibbosa.
A fourth clade includes A. cardenai and is separated from the rest of Iberian species.
The phylogeny of the ant genus Aphaenogaster in the Iberian Peninsula has been recently revised (Lorite et al 2017).Six species present in the Iberian Peninsula (A. iberica Emery, A. senilis Mayr, A. gibbosa, A. subterranea (Latreille), A. dulciniae Emery and A. cardenai) and one Mediterranean species not present (A.spinosa Emery) were included in that analysis.We have analyzed the same seven species, and added A. ulibeli and A. striativentris.In our analysis we have also included Messor capitatus (Latreille).Our results are basically coincidental with those of Lorite et al (2017: Fig 3), repositioning and expanding the A. gibbosa group.

Some additional remarks to these results
Our phylogenetic analysis of COI sequences supports the definition of A. ulibeli as a new species different from the other Iberian Aphaenogaster species and divergent from A. gibbosa.
Looking deeper into this gibbosa-group, the position of A. striativentris is somewhat surprising.Its polymorphism and mandibular morphology could make us think that this species is closely related to the genus Messor.Our result, on the contrary, suggests that it is clearly imbricated into the genus Aphaenogaster, with the mandibular shape being result of convergent evolution.
Another interesting result is the position of A. cardenai, which is excluded from those three main clades and behaves like an outgroup to the genus Aphaenogaster.Its position basal to both Aphaenogaster and the Messor representative (M.capitatus) included in our study suggests that the Aphaenogaster-Messor genera structure is far from being solved.More analysis are needed to define it and they should comprise the Palaearctic Stenammini to map their real affinities.

General Ecology
This species has been found in a Mediterranean mixed forest (Castanea sativa Mill., Quercus robur L.), nesting on ground.
Alated queens and males were found into nests early in June, and males were captured in pitfall traps late in July (27 th -29 th ), suggesting that nuptial flights occur during summer.

Differentiation between A. gibbosa and A.ulibeli
Regarding the worker and queen castes, both species are the only Iberian Aphaenogaster that share the combination

Keys to Iberian Aphaenogaster workers
All images in this key modified from originals at www. antweb.org,links and code numbers under the images.

Fig 1 .
Fig 1 .Definition of ML (A) and MW (B) measurement in the queen caste.

Fig 6 .
Fig 6.ML phylogenetic tree for the COI sequences obtained for species of the ant genus Aphaenogaster in the Iberian Peninsula.A. splendida material was not available, and therefore is not included.A. spinosa, from continental Italy, is also included in the analysis.M. rugiventris is used as outgroup.The tree is drawn to scale, with branch lengths measured in the number of substitutions per site site (bar corresponds to 0.02 substitutions per site).Bootstrap values are indicated on branches when higher than 50.Three of the four clades described in the text are indicated.

Table 1 .
List of ant species and colonies whose COI sequence has been obtained in this study along with voucher specimen codes and relevant sampling data.

Table 4 .
Main differences between A. ulibeli and A. gibbosa males. Supplementary