Updating salamander datasets with phenotypic and stomach content information for two mainland Speleomantes

European plethodontid salamanders (genus Speleomantes; formerly Hydromantes) are a group of eight strictly protected amphibian species which are sensitive to human-induced environmental changes. Long-term monitoring is highly recommended to evaluate their status and to assess potential threats. Here we used two low-impact methodologies to build up a large dataset on two mainland Speleomantes species (S. strinatii and S. ambrosii), which represents an update to two previously published datasets, but also includes several new populations. Specifically, we provide a set of 851 high quality images and a table gathering stomach contents recognized from 560 salamanders. This dataset offers the opportunity to analyse phenotypic traits and stomach contents of eight populations belonging to two Speleomantes species. Furthermore, the data collection performed over different periods allows to expand the potential analyses through a wide temporal scale, allowing long-term studies.


Background & Summary
European cave salamanders are a group of eight amphibians endemic to Italy and to a small part of the French Provence 1 , all belonging to the genus Speleomantes 1 (formerly considered Hydromantes). Three species (S. strinatii, S. ambrosii and S. italicus) are distributed along the Northern and Central Apennines (only S. strinatii naturally extends its range also in France), while the other five (S. flavus, S. supramontis, S. imperialis, S. sarrabusensis and S. genei) are endemic to Sardinia 1 . European cave salamanders are fully terrestrial and lack lungs (Lanza et al., 2006). These features force Speleomantes to select only specific microclimatic conditions: they need high moisture and relatively cold temperatures to survive 2 . Therefore, Speleomantes often inhabit subterranean environments 3,4 , where their preferred microclimatic conditions are realized 5 . Nonetheless, subterranean environments also may be chosen by Speleomantes because predator pressure is lower if compared to epigean ones 6,7 . Indeed, Speleomantes likely represent one of the apex predators in these environments 8 , preying on a wide array of taxa 9 . Speleomantes' narrow eco-physiological requirements, combined with their limited distributions and high site fidelity 1,10 , make these species very sensitive to human-induced effects and susceptible to extinction 11,12 ; all Speleomantes species are therefore strictly protected 13,14 .
Improving our knowledge of species at risk of extinction is fundamental to assess their potential threats and to guarantee their survival 15, 16 . For example, prolonged monitoring may help to understand the impact of specific environmental changes [17][18][19] , allowing to forecast future scenarios and promptly act to protect endangered species [20][21][22] . In this regard, the production of comparable datasets through time is of key importance 9,23,24 . However, data collection may not always be an easy task. Species can occur in habitats that pose challenges to human 2 Scientific Data | www.nature.com/scientificdata www.nature.com/scientificdata/ exploration, as for subterranean environments, where sampling and progression require considerable effort and specific technical skills [25][26][27] . Subterranean habitats are not only hard to find or explore, but their peculiar environmental conditions (e.g., cold temperatures, high moisture, narrow space) might challenge surveyor's stamina with a negative impact on the data recorded 25,28 . Another limit to data collection can be determined by the techniques used to collect information. For example, the old-fashioned research methods involving the sacrifice/harm of individuals are now widely condemned and avoided, especially when concerning protected species 29,30 . Therefore, there is an increasing trend in the use of new harmless alternative approaches 31,32 .
We here describe a new database reporting data on two endangered Speleomantes species that can be handled only under specific national authorizations (see Acknowledgments). This dataset includes information on the population structure, phenotypic traits and diet of individuals belonging to two mainland Speleomantes species: S. strinatii and S. ambrosii. The information gathered here was collected adopting methodologies that limit negative impact on individuals, and can be combined with the two previously published datasets on these species 9,23 , extending the available information in space and time. In this work we gathered data from new populations to cover more area of the species' range, but we also repeated the surveys in previously visited populations, thus providing temporal series of information allowing long-term studies focusing on populations but also on single individuals as well 33,34 . Specifically, we here describe a dataset composed of two types of data: images and stomach contents. High quality images allow extrapolation of data on multiple phenotypic traits (e.g., size, morphology, coloration), obtaining information on the overall population traits and on single individuals as well 23,33,35 . In the era of digitization 36,37 , this new "living" digital catalogue (i.e., collection composed by photos of living organisms) will partially replace the natural history museum collections, becoming, at least for some animal groups, an alternative that not only spare animals lives, but also overcome classic limits such as space needed to store specimens and readiness to be used by the worldwide scientific community with no costs 37 . Another advantage of "living" collections is its repeatability, namely the possibility that individuals can be digitally collected multiple times, allowing to perform long term studies on single individuals and populations as well. The data related to stomach contents can be analysed to study the species' trophic niche and the multiple related traits 34,38,39 . Nonetheless, comparing datasets produced over different time allows to assess potential variation affecting specific species traits and infer on the possible causes 40,41 .

Methods
Surveyed sites. We surveyed eight subterranean sites, three artificial mines and five natural caves (Fig. 1); all these sites fall within the species' natural range 1 . Surveys were performed in 2020, between 9 am and 6 pm during warm and sunny days, periods in which subterranean abundance of Speleomantes is the highest 42 . All sites were surveyed in July, while for six of them surveys were also repeated in September (Table 1). We performed extensive research of Speleomantes within the subterranean sites 43 , covering areas where the exploration was possible without speleological equipment. Salamanders were captured and placed in drilled plastic boxes waiting to be checked. When salamanders sampling was finished, we proceeded with the data collection following this order: (i) assessment of the presence/absence of the mental gland for the identification of adult males (see Fig. 1a in 35 ); (ii) record of body weight using a digital scale (precision 0.01 g); (iii) stomach flushing (see below); (iv) photo shooting (see below). Salamanders were then released in their collection points. www.nature.com/scientificdata www.nature.com/scientificdata/ Stomach flushing. Stomach flushing is a technique enabling to inspect amphibians stomach contents without harming individuals 29,44 . A detailed description of this methodology is provided in Lunghi, et al. 9 . Preserved stomach contents were examined in the lab using an optical microscope, and undigested prey items (or part of them; see 9 ) were recognized at the order level; Staphylinidae (Coleoptera) and Formicidae (Hymenoptera) insect families were considered separately. When possible, also arthropods' different stages where considered as independent prey category. Stomach contents were considered: "empty", if no prey item was observed; "not-identifiable", if the advanced stage of digestion prevented the identification to the order level; "full", if at least one prey item was recognizable. Recognized prey were counted following the method described in Lunghi, et al. 9 . Prey items were rarely integer and the identification was often based on fragments. This condition hampered any potential standardized measurement of prey volume using representative geometric polygons. Stomach flushing was performed on a subsample of the captured salamanders (Table 1). photo shooting. In a dark area of the subterranean environment, captured salamanders were dorsally photographed inside a white soft-box to obtain an homogeneous illumination of the subject and reduce shadows 23 . Before each photo session, a photo was shot to a Pantone colour card X-Rite Colorcheker Passport 2 placed into the soft-box to correctly calibrate the colours and light of photos during post-production 23 . We then photographed salamanders next to a plastic ruler to have a standardize size reference. Please refer to 23 for additional information on this method.

Data Records
The dataset (Photos and stomach contents of two mainland Italian Speleomantes salamanders: data from summer 2020 45

technical Validation
This dataset provides data on two strictly protected amphibian species 13 . Salamanders were sampled following protocols aiming to avoid the spread of potential pathogens 46 ; specifically, we used disposable gloves and disinfected with bleach equipment and boots before changing location. During each month, all surveys were performed within 4 days to limit the variation of environmental conditions which may alter the local ecological opportunity 38,41 . To limit pseudoreplication, each site was surveyed only once per month. On the other hand, the www.nature.com/scientificdata www.nature.com/scientificdata/ two surveys performed on the same populations during different months (July and September) create the condition to test additional hypotheses, like the assessment of temporal variability on salamanders' trophic niche 41 , or even to employ specific software to individually recognize salamanders 47,48 and focus future research on single individuals as well. A blinded stomach contents analysis was performed to limit possible bias 49 . The methodology used to shoot photos enables the production of standardised high quality images with low impact on the species 23,50 . The white calibration before each session avoided potential divergence in light condition and thus, providing standardised pictures. This method allows the creation of a "living" digital collection of Speleomantes 23 , a method that does not only avoid the sacrifice of animals (and thus the related stochastic effects conditioning the evolution of populations) but it is also repeatable. For future updates we will try to include an estimation of prey volume, preferably using methodologies allowing to directly measure the volume (i.e., immerging residuals in a liquid).

Usage Notes
The first part of this dataset is composed of high quality images of individuals of Speleomantes strinatii and S. ambrosii from dorsal view. We suggest the use of the program ImageJ to extrapolate salamanders morphometrics and to estimate their snout-vent length 35 , a fundamental parameter to distinguish juveniles from adults 1 . The images can be also used in R environment (http://www.R-project.org/) to perform analysis on multiple phenotypic traits 51,52 . Considering that the dorsal pattern of Speleomantes does not change through the time 33 , it can be used as natural mark to individually recognize salamanders 48,53 . The repeated surveys performed on the same locations were thought to test the efficacy of specific software to automatically recognize Speleomantes salamanders 47,53 . It has been observed that the ventral pattern of the mainland S. strinatii can be used to individually recognize salamanders, and that software can be employed to do it automatically 54 . However, only the three mainland species have visible pigments on their ventral side 1 , thus alternative methods are needed for analyses embracing all Speleomantes (i.e. including the five Sardinian species). Furthermore, the recognition of Speleomantes using their dorsal pattern limits individuals' handling, a potential source of both stress and pathogens 50,[55][56][57] . The population S_strinatii2 in the previously published dataset 23 included two nearby sites, the S_strinatii2 and S_strinatii5 shown here. To combine these data, the actual S_strinatii2 contains the previous 20 photos (1045074-1045113) and the actual S_strinatii5 the other 20 (1045114-1045142). Furthermore, also the population S_strinatii3 in the previous dataset 23 includes 2 different nearby sites; we therefore suggest to split it in S_strinatii3 with the first 33 photos (1045032-1045056) and in S_strinatii7 with the other 10 (1045057-1045069).
The second part of the dataset is provided in CSV format and is ready to be analysed with R. Populations are codified following 23 . To combine the data on stomach contents with the previous dataset: 9 S_ambrosii2 = Cave_ ambrosii1, S_ambrosii3 = Cave_ambrosii3, S_ambrosii4 = Cave_ambrosii2. The data on stomach contents allows to assess different characteristics of the populations trophic niche 34,38,41 . However, only after verifying whether the single individuals were captured during the two different surveys, it is possible to evaluate the variation of individuals' trophic niche over time.

Code availability
No code was used in this study.