Nesting biology and phylogenetic relationships of the parasitoid-hunting wasp Lindenius pygmaeus armatus (Vander Linden, 1829) (Hymenoptera: Crabronidae)

Nests of Lindenius pygmaeus armatus were examined in northern Poland in Kowalewo Pomorskie and Sierakowo. Adults were encountered from late May to late July. The nests were built in sandy areas and wasteland. Seven nests were observed, of which two were dug up and their structure was examined. The channel was approximately 2.5 mm in diameter and 8–10 cm in the length. The material removed during digging was placed near the nest entrance. The main burrow led to 3–5 cells. The cocoons were approximately 5–7 mm long and 2.5–3.5 mm wide. Females of L. p. armatus provided their nest cells with chalcid wasps averaging 14 prey items per cell. Parasitoids Myrmosa atra and kleptoparasites Senotainia conica were observed entering the burrows. Both females and males of L. p. armatus were detected on the flowers of Achillea millefolium, Peucedanum oreoselinum, Daucus carota, and Tanacetum vulgare. The article also includes phylogenetic relationships of Western Palearctic Lindenius species.

the prey, and it is deposited at the base of the head near one of its limbs 10 . A detailed account of male behavior and intraspecific interactions in aggregations of Lindenius was presented by Miller and Kurczewski 12 . A few years later, same authors published observations of the nesting behavior of Nearctic species L. armaticeps (Fox, 1895), L. buccadentis Mickel, 1916 and L. columbianus errans (Fox, 1895), as well as a review of the world literature on the ethology of the Palearctic L. albilabris, L. panzeri, and L. pygmaeus 4 . The larva of L. pygmaeus was described by Grandi 8,11 , while L. albilabris-by Olszewski et al. 13 . An overview of nesting habits was presented by Kazenas 14 .
The genus Lindenius is also a good example of a group that poses taxonomic identification challenges. The problem with identifying species using the available keys- Kohl 15 , Beaumont 16 , Leclercq 17 and the most recent study by Bitsch and Leclercq 18 containing only 12 species-represents a real necessity to add to the knowledge of the taxonomy of this genus in its various aspects. None of the above-mentioned works allows the identification of all species of this genus. And yet any study of species behavior, ecology or phylogeny must begin with species identification. 19 Taxonomic identification of species is an important issue in biological research, because uncertain taxonomy of a species hinders research on its biology. Limitations in the description and identification of new species characterize a "taxonomic obstacle" that may be overcome using DNA-based methods, such as molecular barcoding 20,21 , phylogeny reconstruction 22,23 and species delimitation [24][25][26] .
The purpose of this study is to supplement the existing information on the nesting biology of L. p. armatus, including (1) nesting behavior of females, (2) activity of females when bringing prey, (3) nest structure, (4) prey range, (5) accompanying parasitoids and kleptoparasites, (6) DNA barcode based identification of the collected material at the study site, and (7) preliminary reconstruction assessment of taxonomic issues related to a possible presence of cryptic species among the Western Palearctic Lindenius species, using the reconstruction of phylogeny and species delimitation, based on the DNA barcode data available to the authors.

Results
Observations of behavior. The nests of Lindenius pygmaeus armatus were found at two sites. The first one was the area of an esker and a sandy road in Sierakowo. The vegetation in the vicinity of this site was dominated by Achillea millefolium L., Artemisia vulgaris L., Berteroa incana (L.) DC., Cerastium holosteoides Fr., Daucus carota L., Geranium pusillum L., Lactuca serriola L., Peucedanum oreoselinumL., Potentilla anserina L., Tanacetum vulgare L., Taraxacum officinale F.H. Wigg., and Trifolium arvense L. The nests were located close to each other (2 nests per 20 cm 2 ) in non-vegetated areas. The second site with nests was a wasteland in Kowalewo Pomorskie among numerous nesting bees Halictus tumulorum (Linnaeus, 1758). The most dominant plants near the nests were the majority same plant species except for Peucedanum oreoselinum L.
No copulation was observed. The process of digging the nest by the female lasted about 55 min (Fig. 1D, Sierakowo site). The tunnel of the burrow following the entrance was almost vertical to a depth of 5 cm, having then slightly diagonal position (at an angle of 10° to 20°, with a diameter of 2.5 mm); its total depth was about 10 cm. The cocoons were approximately 5-7 mm long and 2.5-3.5 mm wide. The burrow ended with a single cell, with other of them located in close proximity (Fig. 1A). The first nest contained 3 cells: one cell was provisioned with 14 prey items from the family of Pteromalidae, another one possessed a cocoon with an adult larva of L. p. armatus (a cocoon consisted of silk threads covered with fragments of chalcid wasps; Fig. 2B,C) and the third one was found with a pupa of a kleptoparasitic fly Senotainia conica (Fallén, 1810) (Diptera: Sarcophagidae). The www.nature.com/scientificreports/ second nest contained five cells with cocoons containing adult larvae of L. p. armatus (Fig. 2D). The excavated soil (while digging) formed a small mound around the nest entrance ( Fig. 1B,C). The entrance to the nest was open during the entire provisioning period. No hunting females were observed, but they supported their prey in flight with their middle and hind legs (Fig. 3B,C). The females were extremely careful when entering the nest, even if there were no kleptoparasites inside: they approached the nest only after making many sudden and very winding detours, after which they had dived directly inside. The time of observation of females bringing prey to the nest was from 9.30 to 18.30. The frequency of provisioning the nest varied from 3 min to 15 min (usually about 5 min). When it was cloudy, the female waited out unfavorable conditions inside the nest. The last observation of the female with prey was on 8 August 2022. All prey items appeared to belong to nine species of Pteromalidae ( Fig. 2A, Table 1). Based on field observations, the larva eats the provisions in about 5 days. During the absence of the female, the parasitoid Myrmosa atra Panzer, 1801 (Hymenoptera: Myrmosidae) was detected twice, and after the arrival of the female with its prey, the kleptoparasitic fly Senotainia conica entered the nest (Fig. 3A).
Adult wasps feed on flowers of the following plant species: Achillea millefolium, Peucedanum oreoselinum, Daucus carota, and Tanacetum vulgare.

Phylogenetic relationships of Western Palearctic species.
A preliminary assessment of phylogenetic relationships of the studied wasp species, using the COI barcode sequences available to us, demonstrated that Lindenius pygmaeus is a sister clade of L. panzeri, a species provisioning the brood cells with Diptera. On the other hand, topology needs a taxonomic revision. Three of ten Western Palearctic representatives of the genus (L. albilabris, L. laevis A. Costa, 1867, and L. pygmaeus) were not delimited correctly as species by the delimitation algorithms GMYC, bPTP and ASAP (Fig. 4). Thus, our analysis demonstrated an ongoing speciation and possible recent divergence of cryptic species in this genus.

Discussion
Fossil records of Crabronini are scarce and are known only from the Tertiary, extending from the Middle Eocene (Baltic amber species) to the Miocene (Dominican amber) 27 .
Significant progress has been made over the past decades in clarifying the phylogenetic relationships of the main Apoidea lineages 1,28 . The results of the research conducted by Sann et al. 1 confirm the monophyly of each of the species-rich subfamilies: Astatinae, Bembicinae and Heterogynaidae, Crabroninae and Dinetinae, Philanthinae and the family Sphecidae. Phylogenetic analysis of the genus Lindenius has not yet been undertaken. L. p. armatus nests are mainly found in sandy, sandy-clay, and loess (flat or slightly sloping) areas [8][9][10][11] , but they were also detected on soda ash postindustrial dumping grounds 31 , under high power lines 32 , and at a former airport 33 . Our results confirm the plasticity in the choice of nesting substrate as nests were observed both on heavily hardened soil (wastelands) and in sandy areas. During the study of the structure of L. panzeri nest, Abrahamsen 34 found that the length and angle of inclination of the main tubule depended on the hardness of the substrate. Nests in all studied species are rarely built singly, usually a few at a time or in small colonies 13,35 . The species studied so far build nests at a depth of 3-12 cm in sand, fine gravel, loess or chalk 4 . At least 95% of the prey of L. pygmaeus armatus were Chalcidoidea and Ichneumonoidea (Hymenoptera), although females occasionally hunted small flies of Ceratopogonidae and Sciaridae (Diptera) 4 . The family most often hunted by L. pygmaeus armatus is Pteromalidae, followed by Eulophidae (Hymenoptera) 4 . The homogeneity of the food base in the conducted research (all belonging to Pteromalidae) illustrates this relationship well: for the first time, we detected two genera (Cyrtogaster Walker, 1833; Psychophagus Mayr, 1904) and five species: Cyclogastrella simplex  It is also worth noting that almost all prey specimens were represented by females. Minkiewicz 10 , and Grandi 8 also found that the majority of chalcidoids were females. Also, the pteromalid species hunted by L. pygmaeus are common taxa 36 . The evolution of nest seeking by Lindenius females has been attributed to factors such as the ability to form dense clusters, the tendency of related wasp species to initiate nests in pre-existing cavities, and a preference for nesting in heavily compacted soil. The search for nests of the same species by males and females may be an effective response to the pressure of kleptoparasitic flies 12 .
The behavior of both males and females can lead to intraspecific interactions. Interactions between females of the Nearctic species L. columbianus (Kohl, 1892) include entrance fights, attempted nest appropriation, prolonged nest visits, and temporary cooperation 12 . In the case of Palearctic species, Minkiewicz 10 repeatedly observed a female of L. panzeri placing her prey in an adjacent nest if the entrance to her own nest was blocked.
In addition, the males of L. panzeri were observed waiting for females at the nest to perform copulation 37 . Unfortunately, no similar behavior was observed in L. p. armatus in this study.
The intraspecific interactions are certainly not fully understood and require thorough research, especially in the case of species nesting in clusters. Interactions in assemblages of this species may be an example of the initial stages of social evolution, developing through the mating of individuals of one generation 12 . Due to the very wide range of preferred prey (Hymenoptera and Diptera) of L. p. armatus and its nesting habitats, the status and the variation of this subspecies needs further focused exploration. In particular, a molecular study of specimens collected from different habitats in different parts of Europe is recommended.
Our preliminary species delimitation exercise supports the hypothesis that L. p. armatus may be considered an independent species. Although this hypothesis must be tested using a larger material and additional DNA markers, it is clear that the data on ecology and reproductive biology of L. p. pygmaeus and L. p. armatus must be analyzed separately.
Considering the probability of ongoing speciation among European Lindenius species, revealed by our phylogenetic analysis, the published data on wide ranges of prey, such as both Hemiptera and Diptera in L. albilabris, or Hymenoptera and Diptera in L. pygmaeus, must be critically reanalyzed. It is possible that the available data must be re-assorted among the undescribed yet cryptic species, which may actually have more specialised niches and a narrower prey choice than previously thought. Our study demonstrates that even in the central Europe with a comparatively very well studied fauna the ethological and ecological observations of insects must be accompanied by the molecular identification of species.

Methods
Field observations. Observations of nesting Lindenius pygmaeus armatus were carried out on the wasteland in Kowalewo Pomorskie (53°10'05.7" N; 18°52'15.5" E) and Sierakowo (53°10'18.9" N; 18°52'21.7" E). The research was conducted from early June to mid-August (2020-2022) on sunny and warm days with a temperature of at least 18 °C. The nests were examined by digging at the stage when the female stopped bringing prey to the nest. While the nest was being examined, its structure was drawn on graph paper. A kleptoparasitic fly pupa found in the cell was placed in an Eppendorf tube and kept until the imago emerged. The pteromalid prey found in the nests was placed in plastic tubes with 70% ethyl alcohol and then identified. The behavior of females, nest digging and provisioning of cells were analyzed using video recordings taken with a Canon CCD-V800E 10X camera and direct observations. Photographs were taken using additional Raynox M-250 macro lens.
Phylogenetic reconstruction. For the assessment of phylogenetic relationships of the study object, we downloaded the mitochondrial DNA cytochrome c oxidase subunit 1 (mtDNA COI) barcodes of all Western Palearctic Lindenius mitotypes available from the Barcode of Life Data System (www. bolds ystems. org) Barcode Index Number (BIN) database 20,21 . The sequences of Oxybelus trispinosus (Fabricius, 1787), Crossocerus assimilis (F. Smith, 1856), and Entomognathus brevis (Vander Linden, 1829) from the same database were used as an outgroup. The dataset was supplemented by a few other Lindenius barcodes, which were obtained according to the methodology of DNA extraction, polymerase chain reaction, and sequencing by Budrys et al. 38 . At the start of analysis, the nucleotide substitution model most likely for the obtained alignment was selected from the list of 24 models available in MEGA 11 22 . Phylogeny testing was proceeded using 10.000 replicates of the bootstrap procedure. Substitution patterns were best described by the general time reversible model with gamma distribution of evolutionary rates, using four discrete categories, and a fraction of invariable sites (GTR+G+I). This model was applied for reconstruction of phylogenetic relationships using the Bayesian inference by means of MrBayes 3.2.7a 23 , and using the Maximum likelihood algorithm, implemented in MEGA 11; for the Bayesian inference, Markov Chain Monte Carlo (MCMC) analyses were run for 1.000.000 generations. For species delimitation, three algorithms were applied: the single threshold General Mixed Yule Coalescent model (GMYC) 24 , implemented as "splits" package in R environment, the Bayesian Poisson Tree Processes method (bPTP) web server (species.h-its.org/ptp/) 25 , and the Assemble Species by Automatic Partitioning (ASAP) web server (bioinfo.mnhn.fr/abi/public/asap/asapweb.html) 26 . The ultrametric dichotomous trees for the GMYC analysis were obtained using the RelTime-ML algorithm, implemented in MEGA 11. The nomenclature of the digger wasps follows 2 , chalcid wasps 39 , plants 40 .

Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.