For our purpose, a record was intended as any report of the finding of one or more individuals of a species in a single location and in a single day.

We searched for all the original records of Geophilidae s.l. in the study area browsing the whole scientific literature reporting records of Chilopoda published up to 2023, however ignoring graduation theses. We also gathered records from the digital catalogues of the major scientific collections of research institutions known to host relevant myriapodological collections and expected to include specimens from the study area: the "Chilobio" centipede collection of the Animal Ecology Group, University of Ljubljana (Ravnjak and Kos 2015); the Bonato-Minelli collection (preserved at the Department of Biology, University of Padova); the collections of the Museo di Storia Naturale di Milano; Museo Civico di Scienze Naturali "E. Caffi" di Bergamo; Museo di Storia Naturale, Verona; Museo Friulano di Storia Naturale, Udine; Naturhistorisches Museum Wien; and Zoologische Staatssammlung München (retrieved from GBIF: Staatliche Naturwissenschaftliche Sammlungen Bayerns (2014)).

The "original source" of a record was intended as the earliest publication reporting the record, if any.

Records were included in the dataset when they were accompanied by at least an indication of the locality (e.g., textual indications, codes of geographical units or geographic coordinates) and a taxonomic identification at the genus level or more precise.

Additional information digitised for each record included any indication, either published in the original source or available from other sources on: time of the recording event (e.g., date, period), habitat (e.g., phytosociological entities, names of plant genera or species type of soil), number and sex of the individuals and the GenBank urls of the available sequences of the main molecular markers employed in molecular taxonomy, phylogeography and population genetics of centipedes, namely the "barcode fragment" of COI, the 16S, 18S and 28S markers, obtained from collected specimens associated with the record. We also queried the BOLD and GenBank databases for additional sequences.

For the name and the structure of the columns of the dataset, we followed the Darwin Core standard (Wieczorek et al. 2012).

For each record, the locality where the animals were found was reported as spelled in the original source (in the column "verbatimLocality). A name of the locality was also provided in the main official language(s) of the country to which the locality belongs (in the column "locality"). These latter names were searched in institutional sources (e.g., websites of local administrative institutions) and in topographic maps (e.g., for Italy, the "Carta Topografica d'Italia" map at the scale 1:25000, available as a Web Map Service at http://wms.pcn.minambiente.it/ogc?map=/ms_ogc/ WMS_v1.3/raster/IGM_25000.map).

For each record, the georeferencing of the locality was reported following the "point-radius" method (Wieczorek et al. 2004) by which the locality is associated with: (i) a coordinate pair (latitude and longitude) of the estimated locality and (ii) a linear distance ("radius"), which indicates the uncertainty of the position. The geographic coordinates were reported either as provided in the original or other sources (if available and reliable) or estimated by us by searching in open-access databases (e.g., OpenStreetMap, https://www.openstreetmap.org), or topographic maps. The uncertainty of the coordinates was estimated for all records: when coordinates were provided, but without their uncertainty, this latter was set to 50 m, a reasonable value for a GPS reading uncertainty in mountain settings and for the precision of topographic maps of the region; when coordinates were not provided, their uncertainty was estimated, based on the textual indication of the locality, also taking into account other available information (e.g., the elevation) and the probable meaning given by the authors of the records to the locality names.

For each record, we reported the taxonomic name used in the original source (in the column "verbatimIdentification"), other names used for the record in subsequent publications (in the column "dynamicProperties"), and the name currently considered valid for the most recent identification (in the column "scientificName").

For the scientificName, we followed the "Checklist of the Italian Fauna" (Bonato and Minelli 2021) as the main source for taxonomy and nomenclature. For species not included in the Checklist (e.g., species only occurring in the non-Italian part of the study area), we mainly followed Bonato and Minelli (2014). However, we provisionally maintained Geophilus carnicus as a valid species, following Peretti and Bonato (2016). We acknowledged Geophilus impressus as the valid name for the species referred to as Geophilus alpinus in the Checklist, following an ICZN opinion (ICZN 2020). We used the name Clinopodes strasseri for a species formerly considered a synonym of Clinopodes carinthiacus and recently recognised as a different species (Peretti et al. 2022). Finally, for the species of Strigamia, we followed the revised and provisional taxonomy and nomenclature by Bonato et al. (2023), and some of the names (Strigamia acuminata, Strigamia carniolensis, and Strigamia crassipes) are intended as species complexes instead of species.

We reported many records of Henia and Geophilus, attributed to four putative undescribed species in their original sources, as identified to the genus level in scientificName. They have been flagged with the provisional labels Henia sp.1, Geophilus sp.1, Geophilus sp. 2. and Geophilus sp.3 in the column "taxonRemarks".

The dataset includes 3293 records, based on about 7700 collected specimens. They are assigned to 39 species or species–species complexes of Geophilidae s.l., of which four putative species are still undescribed.

A total of 1595 records (48%) were already published, while the remaining 1698 are here released for the first time, being only found in unpublished catalogues of scientific collections. The already published records were found in 86 publications since 1847 (a complete list of references is available in Suppl. material 1).

The geographic distribution of the records is heterogeneous (Fig. 2): records are denser in the southern part of the study area, a pattern especially evident for the previously unpublished records.

Figure 2.  

Geographic distribution of the records. Each cell has an area of ~ 40 km². The colour of each cell is associated with the number of records already published (increasing yellow intensity) and to the number of records not yet published (increasing blue intensity). Cells without colour have no records.

For 1830 records (56%), the uncertainty of the geographic coordinates was ≤ 500 m (Fig. 3). Instead, 64 records were not georeferenced since the locality indication did not allow us to estimate the coordinates with an uncertainty lower than 100 km.

Figure 3.  

Frequency distribution of the uncertainty of the position of the records (only records with uncertainty < 100 km were georeferenced and are included in this plot). The inset illustrates the uncertainty of the coordinates of each record (records are arranged from the least to the most accurate on the x-axis; note the logarithmic scale on the y-axis).

The oldest record in the dataset dates to 1847 or before (Koch 1847). In the following period, up to about 1970, approximately 500 other records of Geophilidae centipedes from our study area were published, although, in most of the cases, without a precise indication of the recording year. The rate of collection and release of new records increased significantly during the 1970s and since then, more than 2400 records have been collected (Fig. 4).

Figure 4.  

Frequency distribution of the years when the specimens were recorded. Records with uncertain year were excluded when the uncertainty spanned more than two decades.