An increasing number of ecological studies are incorporating functional traits of organisms to understand global patterns of biodiversity (Díaz et al. 2015), community assembly (Kraft et al. 2008) and ecosystem functioning (de Bello et al. 2010). Functional traits allow us to understand the ecological role of organisms in the community, based on their features and on how the organisms interact in the ecosystem (Nock et al. 2016). Thus, trait based studies allow for the assessment of ecosystem functioning by using species biological traits as a proxy for the functional characteristics of the community assemblage (de Bello et al. 2010). Changes in the functional composition of species assemblages can be related to losses of ecosystem function (Mouillot et al. 2013).

Functional trait databases can be potentially useful for many scientific fields, helping to address questions in systematics, biogeography, macroecology, macroevolution or community assembly. Trait databases are valuable in phylogenetic comparative studies used for the reconstruction of ancestral morphology (Harmon et al. 2010) and to understand how different traits (i.e. body size) have evolved across phylogenetic clades (Kuntner and Coddington 2019), or to link genes with functions or phenotypes (Edmunds et al. 2015). Furthermore, trait-based studies are beneficial in conservation, helping to detect global change impacts or to predict if certain traits are correlated to extinction risk (Chichorro et al. 2019), or even to species invasiveness potential (Nyberg and Wallentinus 2005). Trait variation at intra (Garamszegi and Møller 2010) and interspecific levels (Kazakou et al. 2014) are valuable in comparative approaches or community ecology (Bolnick et al. 2011).

As a measure of the diversity of species niches and their functions, Functional Diversity (FD) has been used to understand how species richness or diversity relates to ecosystem function (Petchey and Gaston 2002, McGill et al. 2006, Petchey and Gaston 2006). It is often seen as reflecting a mechanistic link between taxonomic diversity and ecological processes (Cadotte et al. 2011). Given its implications, the selection of traits to quantify FD must be made with caution. Selected traits should reflect the functioning of species in ecosystems and allow understanding of the connection between FD and the response variable of interest according to individual hypotheses (Nock et al. 2016).

Spiders are among the most diverse and abundant predators in all habitats worldwide, being one of the most species-rich invertebrate orders with more than 48,400 described species (World Spider Catalog 2020). Spiders have been used as models in many ecological and evolutionary studies, including functional diversity studies from global (Cardoso et al. 2011) to regional (Cardoso 2012) scales, and methodological advances in phylogenetic and functional diversity analyses (Cardoso et al. 2014). Under this same framework, we are now building a series of analyses that require trait and FD-based statistics, for which the current compilation of traits will serve as a data standard.

We here present a functional trait database of spiders, selecting morphological and ecological traits related to species life histories, microhabitat and trophic preferences, allowing us to characterize the ecological roles of spiders in biological communities and their use of biotic and abiotic resources.

This dataset contains six morphological traits, four ecological traits and two other traits related to origin and dispersal ability.

The database consists of two sections:

1. individual-level data for 6 morphological traits (total body size, prosoma length, prosoma width, prosoma height, tibia I length and fang length), based on direct measurements of 2844 specimens of 506 spider species; and
2. species-level aggregate data for 12 traits (the same 6 morphological traits as the previous section plus dispersal ability, vertical stratification, circadian activity, foraging strategy, trophic specialization and colonization status), based on either average direct measurements or bibliographic search.

Morphological traits, vertical stratification and nocturnality were extracted from the samples whenever possible. The combination of sampling methods provides us with direct information on the vertical stratum preferred by the species, from lower to higher stratification: pitfall traps (ground-dwelling species), vegetation sweeping (low vegetation species), active aerial searching and foliage beating (arboreal species). The time of day when species were collected (diurnal vs. nocturnal sampling) provides information about their circadian activity, although this can be biased given the methods used and should be used with caution. Species collected only in pitfall traps did not have such data, with information being extracted from the literature or based on our own field experience instead.

For morphological data, specimens were selected randomly from the available field collections and the measurements were made using a binocular stereoscope. All measurements are given in millimeters (mm). Each morphological trait was measured, whenever possible, for 10 individuals of each species, 5 females and 5 males. For some species only one individual of each sex (the specimen with median body size) was selected for measuring morphological traits other than body size.

In the second section, aggregate data at species level were taken as the average (by sex and total average) of the previous measures for morphological data when available, or from the literature when no individuals of a species could be measured. Additionally, the minimum and maximum ranges of body length for each sex were included. For vertical stratification and circadian activity, we calculated an index based on the collection data or, if not available, using the literature or personal knowledge of species. Other measures were taken from the literature, or often from personal experience with the species in the field.

1. Morphological traits (continuous variables) in millimeters (mm):

• Total body length: measurement of the distance between the front of the carapace (without the chelicerae) and the end of the abdomen (without spinnerets).
• Prosoma length: measurement of the maximum carapace length (without the chelicerae).
• Prosoma width: measurement of the maximum carapace width.
• Prosoma height: measurement of the maximum prosoma height (in lateral view).
• Tibia I length: measurement of the retrolateral length of the first tibia.
• Fang length: measurement of the maximum chelicerae fang length.

2. Ecological traits:

• Verticality (continuous variable): information about the vertical stratification (from epigean to arboreal), based on the information provided by the sampling methods with which the species were collected during fieldwork. One value was assigned to each sampling method (from lower to higher level): pitfall (0), sweeping (1), active aerial searching (2), and beating (3). A weighted mean considering the abundance was performed and this was divided by 3 to rescale the values to a range from 0 to 1. Lower values indicate ground-dwelling species, while higher values indicate arboreal species.
• Nocturnality (continuous variable): information about circadian activity (nocturnal/diurnal) of the species. Diurnal species were recorded as 0, nocturnal species as 1, and for species that can be either diurnal or nocturnal we took a value based on the proportion of individuals of each species active at night during sampling (we divided the number of specimens collected during the night by their total abundance). Values ranged from 0 to 1, with lower values indicating diurnal species and higher values indicating nocturnal species.
• Foraging strategy (eight binary variables; 0 no web - 1 web builders): information about the method of hunting or type of web the species builds. For the web-building species: capture web, sensing web, tube web, sheet web, space web and orb web (more than one feature can be true). For hunting species: ambush hunter and active hunter. The classification was usually made at genus level following the criteria described in Cardoso et al. (2011).
• Trophic specialization (binary variable): information about prey range: stenophagy (specialist species recorded as 1), and euryphagy (generalist species recorded as 0). The classification was generally made at genus level following the criteria described in Cardoso et al. (2011).

3. Other traits:

• Colonization status (categorical variable): for the species distributed in the Macaronesian archipelagoes, information about their origin was included: (E) endemic to the archipelago; (N) native non-endemic; and (I) introduced; according to the Canary Island (Macías-Hernández 2010), Madeira (Cardoso and Crespo 2008) and Azores (Cardoso et al. 2010) spider checklists. The origin status of some species from the Canary Islands was modified according to our knowledge of the species distribution and it will be updated in the public database "Banco de Datos de Biodi­versidad de Canarias" (http://www.biodiversidadcanarias.es/atlantis/common/index.jsf).
• Dispersal: information about dispersal ability was based on ballooning propensity (see also Carvalho and Cardoso 2014). Given the data in Bonte et al. 2003, Bell et al. 2005, Blandenier 2009, we considered as frequent (F) ballooners all species from the families that constituted more than 10% of aerial catches in past studies, in this case Linyphiidae and Araneidae. As occasional (O) ballooners those that made up between 1 and 10%. As rarely (R) or never ballooning species those where ballooning has never been detected or usually represents less than 1% of the catches in published ballooning studies.

This study was supported by the project BIODIV ISLAND-CONT (Biodiversity drivers on islands and continents – 706482) funded by Marie Sklodowska-Curie Individual Fellowships (H2020-MSCA-IF-2015) to the first author NMH. The research was additionally funded by three other projects that provided the material collected for the database:

1. the ERA-Net Net-Biome research framework, financed through Portuguese FCT-NETBIOME grant 0003/2011 (PB);
2. FCT MACDIV – FCT-PTDC/BIABIC/0054/2014 (PB, BE, PC);
3. the Organismo Autónomo de Parques Nacionales Spain (OAPN #485/2012) (MA).

Additional support was provided by grant 2017SGR83 from the Catalan Government to MA. PB and PC are supported by AZORESBIOPORTAL–PORBIOTA (ACORES-01-0145-FEDER-000072) for the development of Functional Traits for Azorean arthropods. MD is supported by an APIF PhD fellowship from the University of Barcelona. We also thank Rutger Vos and Martin Ramirez for their valuable comments that greatly helped to improve the quality of this paper.

CR, MD, SF and IS measured the specimens. PC collected information about ecological traits at genus level. NMH wrote the manuscript with support from PC. CR made the tables. All authors commented on different versions of the manuscript.