A new species of yellow acorn ant discovered in Italy via integrative taxonomy ( Temnothorax luteus -complex, Formicidae)

The Mediterranean, a global hotspot for rare ant species, hosts a significant representation of the global diversity of the mainly Holarc - tic ant genus Temnothorax . However, several groups still require significant taxonomic efforts. The taxonomy of the T. luteus complex species was revised in 2014 when morphometrics allowed distinguishing two valid species and two synonyms out of four taxa that had been originally described from France. The two species recognized since then are T. luteus , distributed from Iberia to the Alps, and the largely sympatric but much more xerothermophilic T. racovitzai . In Italy, only a few records of the complex were ever published, and the identity of the Italian population was never thoroughly assessed. We combined morphometrics with phylogenomic data to assess the identity of the T. luteus populations that spanned from Sicily to the Italian Alps and discovered that all Italian samples belong to a new cryptic species, which we describe as T. apenninicus sp. nov. whose glacial refugium was probably in the southern Apennines.


Introduction
The ant genus Temnothorax Mayr, 1861 (Formicidae, Myrmicinae) is mainly distributed across the Holarctic region and northern Neotropics (with a minor presence in the northern Afrotropics) and currently counts about 465 extant species (Prebus 2015(Prebus , 2017;;Bolton 2024).Temnothorax ants are generally characterized by small-sized workers and small colonies that often nest inside microspaces such as acorns, rock crevices, empty gastropod shells, or galls (thus earning the name of acorn ants) (Prebus 2017;Seifert 2018;Giannetti et al. 2022aGiannetti et al. , 2022b)).While easily overlooked during field surveys, these ants may play an essential ecological role by preying upon a range of small arthropods (Seifert 2018;Giannetti et al. 2022aGiannetti et al. , 2022c)).
The description of new Temnothorax species has sped significantly during the last few decades (Bolton 2024), while a global phylogenomic framework is now available (Prebus 2017).The Mediterranean region, which is a global hotspot for rare ants (Kass et al. 2022), harbors a high number of narrowly distributed species of Temnothorax, whose biogeographic patterns often mirror the intricate paleogeographic and climatic history of the region (Schifani et al. 2022;Wang et al. 2023).
The Temnothorax luteus complex of yellowish ground-nesting species has been revised by Seifert et al. (2014), who, using morphometric data, demonstrated it to comprise two separate cryptic species in the western Mediterranean, namely T. luteus (Forel, 1897) and T. racovitzai (Bondroit, 1918), originally described from France alongside with their synonyms.T. luteus was demonstrated to be a senior synonym of T. tristis (Bondroit, 1918), and T. racovitzai a senior synonym of T. massiliensis (Bondroit, 1918), as the previous criteria used to delimit and identify the four taxa proved to be inconsistent (Seifert et al. 2014).The T. luteus complex belongs to the rottenbergii clade alongside the T. exilis and T. rottenbergii groups (Schifani et al. 2022).Temnothorax luteus and T. racovitzai have partly overlapping distributions but are characterized by different ecological preferences: T. racovitzai occurs in coastal, xerothermic and relatively open habitats of France and Iberia, which is widespread and common.On the other hand, T. luteus ranges from the Iberian Peninsula to the Alps, north to at least 46.6, prefers overall colder sites and higher elevation, and in Iberia, its distribution is more limited (Seifert et al. 2014;Seifert 2018;Arcos and García 2023).Seifert et al. (2014) hypothesized that the two species evolved in two different glacial refugia, namely the Italian peninsula (T.luteus) and the Iberian peninsula (T.racovitzai), which host significantly different ant faunas and acted as separate glacial refugia (Schifani 2022;Arcos and García 2023;Wang et al. 2023).However, at that time, no samples from Italy, where the complex has rarely been recorded (Baroni Urbani 1971), were available for investigation (Seifert et al. 2014).
We investigated the populations of the T. luteus group occurring in Italy by integrating the morphological and phylogenetic data.

Material examined
This study analyzes the morphometric data of 221 worker individuals belonging to 82 nest samples.Forty worker individuals belonging to 18 nest samples were measured by SC, extending the morphometric dataset of Temnothorax luteus complex species published by Seifert et al. 2014 following the same morphometric protocol.In addition, SC measured eight individuals of 3 nest samples that were measured by BS in 2014 to test the reproducibility of the morphometric measurements of the two gaugers (Csősz et al. 2021).The complete list of examined material is listed in Suppl.material 1, including locality information, number of measured individuals.

Protocol for morphometric character recording
All measurements were made with a cross-scale graticule at µm precision using a pin-holding stage, permitting rotations around X, Y, and Z axes with an Olympus SZX16 stereomicroscope with a 1.6× Plan Apochromat objective at a magnification of ×192 for each character by SC.Definitions of morphometric characters are listed in Table 1.Raw data in millimeters are given in Suppl.material 2. We used intraclass correlation coefficients (ICC) to calculate the reproducibility parameter between the dataset measured by BS and the extension added anew by SC via package ICC (Wolak et al. 2012) in R (see Table 1).

Statistical framework on morphometric data-hypothesis formation and testing
We use the toolkit of exploratory data analysis of continuous morphometric data (Seifert et al. 2014;Csősz and Fisher 2016) followed by confirmatory data analysis.

Exploratory analyses via NC-PART clustering
The prior species hypothesis was generated based on workers via Nest Centroid clustering (NC clustering; Seifert et al. 2014) in combination with partitioning algorithms PART (Nilsen and Lingjaerde 2013;Csősz and Fisher 2016) for estimating the number of biologically meaningful clusters.The protocol of this combination was published by Csősz and Fisher (2016), which is now applied with the following specific setups: bootstrap iterations in PART were set to 'b=1000', and the minimum size of clusters was set to 'minSize=5' for both 'hclust' and 'kmeans'.Data input for this analysis was performed as absolute measurements (raw data).

Confirmatory data analyses
The validity of the prior species hypothesis imposed by the exploratory processes was tested via a cross-validated linear discriminant analysis (CV-LDA) using the package MASS (Venables and Ripley 2002).Statistical analyses have been done in R (R Core Team 2023).Conventional LDA and backward stepwise methods were used to create an easy-to-use numeric key for separating species via character reduction.Data input for this analysis was performed as absolute measurements (raw data).

Taxon sampling
We gathered three specimens of the new species from two localities and incorporated these with three sequences from a previously published study (Schifani et al. 2022; see Suppl.material 3 for collection data).

Molecular data generation
To generate the genetic data, we extracted DNA, prepared genomic libraries, and performed targeted enrichment of ultraconserved elements (UCEs).We extracted DNA nondestructively from adult worker ants by using a flame-sterilized size 2 stainless steel insect pin to pierce the cuticle of the head, mesosoma, and gaster on the right side of the specimens, then used a DNeasy Blood and Tissue Kit (Qiagen, Inc., Hilden, Germany) following the manufacturer's protocols.We verified DNA extract concentration using a Qubit 3.0 Fluorometer (Invitrogen, Waltham, MA, U.S.A.).We input up to 50 ng of DNA, sheared to a target fragment size of 400-600 bp with a QSonica Q800R sonicator (Qsonica, Newtown, CT, U.S.A.), into a genomic DNA library preparation protocol (KAPA HyperPrep Kit, KAPA Biosystems, Wilmington, MA, U.S.A.).For targeted enrichment of UCEs, we followed the protocol of Faircloth et al. (2015) as modified by Branstetter et al. (2017) using a unique combination of iTru barcoding adapters (Glenn et al. 2019; BadDNA, Athens, GA, U.S.A.) for each sample.We performed enrichments on pooled, barcoded libraries using the catalog version of the Hym 2.5Kv2A ant-specific RNA probes (Branstetter et al. 2017; Arbor Biosciences, Ann Arbor, MI, U.S.A.), which targets 2,524 UCE loci in the Formicidae.We followed the library enrichment procedures for the probe kit, using custom adapter blockers instead of the standard blockers (Glenn et al. 2019; BadD-NA, Athens, GA, U.S.A.), and left enriched DNA bound to the streptavidin beads during PCR, as described in Faircloth et al. (2015).Following post-enrichment PCR, we purified the resulting pools using SpeedBead magnetic carboxylate beads (Rohland and Reich 2012;Sigma-Aldrich, St. Louis, MO, U.S.A.) and adjusted their volume to 22 μL.We verified enrichment success and measured size-adjusted DNA concentrations of each pool with qPCR using a SYBR-FASTqPCR kit (Kapa Biosystems, Wilmington, MA, U.S.A.) and a Bio-Rad CFX96 RT-PCR thermal cycler (Bio-Rad Laboratories, Hercules, CA, U.S.A.), subse- The maximum cephalic length in the median line; the head must be carefully tilted to the position with the true maximum.
Excavations of occiput and/or clypeus reduce CL.
0.807 CS Cephalic size, the arithmetic mean of CL and CW, is a less variable indicator of body size.

NA CW
The maximum cephalic width is measured across and including the eyes, exceptionally posteriorly of the eyes.0.966 EYE Eye-size index: the arithmetic mean of the large (EL) and small diameter (EW) of the elliptic compound eye is divided by CS, i.e.EYE=(EL+EW)/(CL+CW).All structurally visible ommatidia are considered. 0.896

FRS
The distance of the frontal carinae immediately caudal of the posterior intersection points between the frontal carinae and the lamellae dorsal of the torulus.If these dorsal lamellae do not laterally surpass the frontal carinae, the deepest point of scape corner pits may be taken as a reference line.These pits take up the inner corner of the scape base when the scape is entirely switched caudad and produce a dark triangular shadow in the lateral frontal lobes immediately posterior to the dorsal lamellae of the scape joint capsule. 0.983

MW
The maximum mesosoma width; pronotal width in workers.0.877 ML The mesosoma length.Measured from the caudal-most point of the propodeal lobe to the transition point between the anterior pronotal slope and anterior propodeal shield (preferentially measured in lateral view; if the transition point is not well defined, use dorsal view and take the center of the dark-shaded borderline between pronotal slope and pronotal shield as an anterior reference point). 0.982

PEH
The maximum petiole height.The straight section of the ventral petiolar profile at node level is the reference line perpendicular to which the maximum height of the petiole node is measured.The maximum straight line scape length excluding the articular condyle as the arithmetic mean of both scapes.0.902 SPBA The smallest distance of the lateral margins of the spines at their base.This should be measured in the dorsofrontal view since the wider parts of the ventral propodeum do not interfere with the measurement in this position.If the lateral margins of spines diverge continuously from the tip to the base, the smallest distance at the base is not defined.In this case, SPBA is measured at the level of the bottom of the interspinal meniscus.
0.966 SPST Distance between the center of propodeal stigma and spine tip.The stigma center refers to the midpoint defined by the outer cuticular ring but not to the center of the real stigma opening which may be positioned eccentrically. 0.641

SPTI
The distance of spine tips in dorsal view; if spine tips are rounded or thick, the centers of spine tips are taken as reference points.

0.997
quently combining all pools into an equimolar final pool.We sequenced the final pool in one lane at Novogene (Sacramento, CA, U.S.A.) on Illumina HiSeq 150 cycle Paired-End Sequencing v4 runs (llumina, San Diego, CA, U.S.A.), along with other enriched libraries for unrelated projects.

Dataset construction
We followed the standard PHYLUCE protocol for processing UCEs in preparation for phylogenomic analysis (Faircloth 2016), aligning the monolithic unaligned FAS-TA file with the phyluce_align_seqcap_align command, using MAFFT (Katoh and Standley 2013) as the aligner (--aligner mafft) and opting not to edge-trim the alignment (-no-trim).We trimmed the resulting alignments with the phyluce_align_get_gblocks_trimmed_align-ments_from_untrimmed command in PHYLUCE, which uses GBlocks ver.0.91b (Castresana 2000), using the following settings: b1 0.5, b2 0.5, b3 12, b4 7.After removing UCE locus information from taxon labels using the command phyluce_align_remove_locus_name_from_ nexus_lines, we examined the alignment statistics using the command phyluce_align_get_align_summary_data and generated a dataset in which each locus contains a minimum of 85% of all taxa using the command phy-luce_align_get_only_loci_with_min_taxa.

Phylogenetic inference
Because the assumption that the evolutionary rates of sequence data are homogenous is often violated in empirical data (Buckley et al. 2001), we partitioned our UCE loci into sets of similarly evolving sites.To achieve this, we used the command phyluce_align_format_nex-us_files_for_raxml which concatenates loci into a single alignment and generates a partition file for input into the SWSC-EN method (Tagliacollo and Lanfear 2018).We used the resulting datablocks as input for partitioning in IQTREE ver.2.1.2(Minh et al. 2020), using the command -m MFP+MERGE.Because the combination of gamma and proportion of invariable sites (+I+G) has been demonstrated to result in anomalies in likelihood estimation (Sullivan & Swofford, 2001;Yang, 2006), we set the rate heterogeneity models to a subset that includes everything except the combination of gamma and proportion of invariable sites (-mrate E, I, G) and set the search algorithm to -rclusterf 10.We used the resulting partitioned dataset as input for maximum likelihood tree inference in IQTREE, using 1000 ultrafast bootstrap replicates (-B 1000).

Morphometry
The two clustering methods 'hclust' and 'kmeans' of PART in combination with NC-clustering resulted in three clusters confirming the validity of the two formerly recognized West European Temnothorax luteus complex species T. luteus and T. racovitzai and a third, Italian cluster took shape.The partitioning methods returned two alternative classifications for two samples, and five more samples have been assigned to outliers out of the total 82 (Fig. 1).These samples with conflicting identifications were set to wild cards in the first run of confirmatory LDA (Seifert et al. 2014).In the second LDA run and the LOOCV-LDA, these samples are labeled according to the posterior probability values the first LDA returned.The LDA yielded an overall accuracy of 0.97 [95% CI: (0.94, 0.99)] in the classification hypothesis in individuals (Fig. 2) and returned 100% accuracy in the nest sample mean level (Fig. 3).
The cross-validation LDA involving all characters confirmed this classification on an individual level with 95.5% probability (overall accuracy: 0.9545, and 95% CI: 0.918, 0.978) (see Table 2).This means a 4.5% error rate (10 misclassifications out of the 220 individuals) in classification.

Molecular data processing
Following assembly and UCE extraction, the mean number of loci per sample was 2365, with a mean contig length of 1279 bases and a mean coverage depth of 46.9× (see Suppl.material 4 for summary statistics).Following alignment, trimming, and filtering the full UCE dataset to loci with ≥80% taxon presence (including samples from previous studies), the final dataset had 2385 loci, with a mean locus alignment length of 1118 bases.The concatenated matrix was 2.7 Mb in length, in which 118 Kb were variable, 12 Kb were parsimony informative, with 13.9% missing data.Raw sequence reads newly generated for this study can be found on the NCBI Sequence Read Archive (BioProject PRJNA1092267).

Phylogenetic inference
The partitioning analysis of the dataset in IQTREE resulted in a 118-partition scheme.Model selection resulted in the assignment of 31 unique substitution models.The resulting tree received full statistical support for each node, with the new Temnothorax species forming a clade separate from T. luteus and T. racovitzai, which were recovered as sister species (see Fig. 4).
Body color yellow; light brown.Body color pattern concolorous except for the posterior margin of the first gastral tergite which is often characterized by a weak transversal band interrupted in its central portion.Appendages yellow, except for the usually infuscate antennal clubs.Queen (Fig. 6 A-C).Body color yellowish brown, with anterior side of gastral tergites yellow, and appendages yellow except the infuscate antennal clubs.Head subrectangular, clypeal depression absent, clypeus smooth and shiny with vertical costae running parallel on the sides and no central carina.Head frontal sculpture: main sculpture costate in the center to areolate-rugose near the posterior and lateral margin, ground sculpture weakly areolate to smooth.Gena sculpture: main sculpture costate, ground sculpture weakly areolate.Mesosoma dorsum flat in profile view, propodeal spines well-developed, horizontal and with a wide base.Dorsal region of mesosoma sculpture: mesoscutum longitudinally costate, mesoscutellar disk largely smooth and shiny, propodeum scabrous.Lateral region of mesosoma sculpture: mostly covered in weak longitudinal costae, converging towards the spine in the propodeum, except for the katepisternum which is mostly smooth and shiny.Frontal profile of petiolar node contour line in lateral view concave.Subpetiolar process well developed, tooth-like.Petiole and postpetiole sculpture: main sculpture scabrous, ground sculpture areolate.Surface of the first gastral tergite smooth.
Male (Fig. 7 A-C).Body color brown, appendages pale yellowish.Head subtriangular, clypeal depression absent, clypeus smooth and shiny with vertical costae running parallel on the sides and no central carina.Head sculpture: ground sculpture conspicuous areolate, main sculpture carinate-rugose near the ocelli.Propodeal spines absent.Dorsal region of mesosoma sculpture: mesoscutum anteriorly smooth and posteriorly finely covered in irregurar longitudinal costae, mesoscutellar disk with fine longitudinal costae, propodeum areolate.Lateral region of mesosoma sculpture: largely areolate, katepisternum mostly smooth and shiny.Frontal profile of petiolar node contour line in lateral view concave.Petiole and postpetiole sculpture: main sculpture areolate, part of the petiolar dorsum and whole postpetiolar dorsum smooth.Surface of the first gastral tergite smooth.
Etymology.The name refers to the Apennines Mountains that span from Sicily to northern Italy, encompassing most of the species' distribution and likely constituting its glacial refugium.
Morphological diagnosis.The only qualitative difference between T. apenninicus and T. luteus is represented by the normally infuscate antennal clubs and infuscate first gastral tergite of the first (features that may be subsequently lost depending on the preservation state of the specimens).In the Italian fauna, a chromatic pattern superficially similar to T. apenninicus is known for T. minozzii (Santschi 1922), which is however characterized by much shorter spines, a different petiole shape and a smooth head sculpture, and only known from a lowland site.
The three species considered in this revision represent otherwise true cryptic species, as no single morphometric ratio allows for separation at the individual level.The nest sample mean values spine length ratio (SPST/CS) help to tell Temnothorax apenninicus sp.nov.workers apart from its congeners T. luteus and T. racovitzai with an acceptably high success rate (Fig. 8, Table 4).However, a narrow overlap is still present.Distribution and ecology.The range extends from Sicily to the Alps (Fig. 11).Temnothorax apenninicus appears to be a mountain species that lives preferably in meadows and open habitats.Nests are built opportunistically on the ground, in rock crevices, or under stones.Most of our data come from the southern portion of T. apenninicus distribution range: T. apenninicus is probably the highest elevation Temnothorax species in Sicily (Etna, Mt.Carbonara), ranging from 1400 to 2000 m.At higher latitudes, we expect a downshift of the elevational range, with the single record from central Italy (Latium) at 1020 m, and the one from the southern edge of the Alps at 1460 m.On the Etna, multiple colonies were found to be parasitized by the social parasite T. muellerianus   (Finzi, 1922) also referred to as Chalepoxenus muellerianus (see Ward et al. 2015Ward et al. , 2016;;Seifert et al. 2016).Buschinger et al. (1988) noted that the local population of this social parasite (previously known as Chalepoxenus siciliensis Kutter, 1973) seemed to have specialized in parasitizing a species "close to Leptothorax tristis" (i.e., a former name of T. luteus), which most likely was T. apenninicus.

Dichotomic key for worker specimens' identification
The W-European representatives of the T. luteus complex represent a morphologically cryptic triad, hence only a combination of morphometric traits can yield safe solutions in separating these species.

Discussion
The discovery of T. apenninicus, stretching from Sicily to the Italian Alps, is paired with the absence of T. luteus and T. racovitzai records in the same range.The easternmost distribution of T. luteus reaches the French and Swiss Alps, and the species is likely present on the western side of the Italian Alps as well (Menchetti et al. unpublished data), so the contact range of the two certainly requires further investigation.However, the few and scattered previous T. luteus records in Italy were most likely based on misidentifications of T. apenninicus if not of other taxa (Baroni Urbani 1971;Schifani et al. 2022Schifani et al. , 2024)).
Our results reinforce the previous hypothesis that T. racovitzai may have colonized France from an Iberian glacial refugium (Seifert et al. 2014), as it does not appear to occur east of France.While the southern Apennines must have been the glacial refugium of T. apenninicus, T. luteus may have been perhaps isolated in the western Alps (Taberlet et al. 1998;Menchetti et al. 2021;Schmitt et al. 2021).While some morphologically more distinct species may belong to the luteus group sensu lato both on the western (e.g., T. pardoi (Tinaut, 1987), see Gouraud et al. 2021) or perhaps even eastern Mediterranean basin (e.g., as suggested for the Cretan T. variabilis Salata, Borowiec & Trichas, 2018-see Salata et al. 2018), no phylogenetic data are yet available to assess this hypothesis and the definition of Temnothorax species groups solely based on qualitative morphological characters remains elusive (Schifani et al. 2022).
Italy hosts several endemic ant species, with no assessment of their conservation status available yet (Schifani 2022).Only very few of these, including T. apenninicus, seem to be restricted to high elevations -including one, Aphaenogaster italica with a distribution very similar to T. apenninicus (Radchenko et al. 2006;Schifani and Alicata 2018).The status of these species should receive attention considering the threat climate change is posing to mountain ecosystems (Guisan et al. 2019;Schifani et al. 2024).While molecular evidence is expanding our understanding of intrageneric relationships in Temnothorax (Prebus 2017), the picture in the Mediterranean is still very incomplete on many fronts, and not all groups are easily delimited on a morphological basis (Schifani et al. 2022).Further investigations will be helpful to identify other potential relatives of the T. luteus complex in neighboring regions, such as Iberia and the Maghreb, and eventually attain a clearer picture of the diversity of the whole rottenbergii clade (Prebus 2017).
in lateral view; measured from the anterior corner of the subpetiolar process to the dorsocaudal corner of the caudal cylinder.0.931 PEW maximum width of petiole.0.988 PoOc The postocular distance.Use a cross-scaled ocular micrometer and adjust the head to the measuring position of CL.Caudal measuring point: median occipital margin; frontal measuring point: median head at the level of the posterior eye margin.Note that many heads are asymmetric and average the left and right postocular distance.

Figure 2 .
Figure 2. Linear Discriminant Analysis (LDA) plot on worker individuals of Temnothorax luteus complex species.The position of Temnothorax apenninicus (blue), T. luteus (dark green), T. racovitzai (purple) individual workers are illustrated in a morphospace defined by LD1 and LD2 scores.

Figure 3 .
Figure 3. Linear Discriminant Analysis (LDA) plot on nest sample means of Temnothorax luteus complex species.The position of Temnothorax apenninicus (blue), T. luteus (dark green), T. racovitzai (purple) nest samples are illustrated in a morphospace defined by LD1 and LD2 scores.

Figure 4 .
Figure 4. Maximum likelihood phylogeny resulting from the partitioned analysis of 2385 ultraconserved element loci.Statistical support at each node is in bootstraps.

Table 1 .
Abbreviations for morphometric characters, character definitions, and the Intraclass Correlation Coefficient as a reproducibility parameter are presented in different columns.