Spatial distribution of Madeira Island Laurisilva endemic spiders (Arachnida: Araneae)

Abstract Madeira island presents a unique spider diversity with a high number of endemic species, many of which are still poorly known. A recent biodiversity survey on the terrestrial arthropods of the native forest, Laurisilva, provided a large set of standardized samples from various patches throughout the island. Out of the fifty two species recorded, approximately 33.3% are Madeiran endemics, many of which had not been collected since their original description. Two new species to science are reported – Ceratinopsis n. sp. and Theridion n. sp. – and the first records of Poeciloneta variegata (Blackwall, 1841) and Tetragnatha intermedia Kulczynski, 1891 are reported for the first time for Madeira island. Considerations on species richness and abundance from different Laurisilva locations are presented, together with distribution maps for endemic species. These results contribute to a better understanding of spider diversity patterns and endemic species distribution in the native forest of Madeira island.


Introduction
Interest on Madeira spider fauna started almost two centuries ago (Blackwall 1859, Blackwall 1862, Bosenberg 1895, Bristowe 1925, Denis 1962, Denis 1963, Denis 1964, Johnson 1863, Kulczynski 1899, Kulczynski 1905, Lowe 1832, Schenkel 1938, Schmitz 1895, Warburton 1892 and has continued until recently, with the description of several new species (Wunderlich 1987, Wunderlich 1992, Wunderlich 1995. Despite the major relevance for the knowledge of Madeiran spider biodiversity, these taxonomic studies are based on traditional ad-hoc manner, without a standardized methodology, which renders them inappropriated to explore species distribution patterns. In other words, although the ad-hoc approach may be useful in targeting specific localities or ecological niches by experienced collectors ), it does not capture the real distribution of species. The lack of knowledge on species distribution is now referred to as the Wallacean shortfall (Lomolino 2004; see Cardoso et al. 2011), and the main objective of this contribution is to help to overcome this shortfall for Madeiran spiders associated with the native laurel forest -Laurisilva.
Laurisilva is a relict evergreen forest that covered part of the Mediterranean Basin during the Tertiary, being restricted nowadays to a few locations in the Macaronesian archipelagoes of the Azores, Madeira and Canary Islands, and in some secluded sites of North Africa (Neves et al. 1996). In the archipelago of Madeira, Laurisilva is restricted to the main island (Madeira), covering about 16000 ha and harbouring a large number of endemic fungi, plants and animals (Uva 2008; . Taking in consideration the high natural value of this forest type, now internationally recognized (IUCN 1999, Myers et al. 2000, local entities created the Natural Park of Madeira with the aim of valuing, studying and protecting this unique ecosystem. The latest spider checklist (Cardoso and Crespo 2008) reports the occurrence of 164 species for Madeira island. Among these species, 40 are considered single-island endemics (SIE) and some belong to genera that radiated in Madeira archipelago (e.g. Dysdera Latreille, 1804, Hogna Simon, 1885, Pholcus Walckenaer, 1805. Approximatelly 63% of the SIEs were collected exclusively from Laurisilva sites, strongly suggesting an intimate association between these species and this type of habitat. Although the Natural Park of Madeira has developed major efforts to protect and restore Laurisilva patches from human activities and invasive species, conservation measures specifically targeting endemic terrestrial arthropod fauna are yet to be put forward. Preliminary conservation priorities for the endemic invertebrate fauna have already been identified using taxonomically unbiased criteria (Martin et al. 2010, Martin et al. 2008), but there is still a lack of baseline information for most species. The collection of species distribution and abundance data using standardized sampling is paramount to assess species conservation status. The aim of our study is therefore to provide information on the distribution of spiders endemic to the Laurisilva of Madeira island based on a large-scale assessment using standardized sampling.

Material and methods
Spiders were sampled using two standardized complementary techniques: pitfall trapping and direct sampling. Pitfall sampling was performed both in May-June 2006 (sites 1 to 28) and June 2007 (sites 29 to 36). A linear transect with 30 pitfall traps, 5 m apart, was set in each study site. Each trap consisted of a plastic cup (4.2 cm diameter by 7.8 cm height) protected with a plastic cover (15 cm diameter) attached~3 cm aboveground to prevent flooding and loss of specimens. Ethyleneglicol (10%) and Turquin solution were used as preservatives in alternate traps along the linear transect (BALA protocol, Borges et al. 2006, Borges et al. 2005. The traps were active during a two-week period. Direct sampling was performed in June 2007 for most sampling sites. The sampling design was semi-quantitative with a sampling unit defined as 30 minute/person of effective fieldwork. Direct sampling included the search of spiders in three microhabitats (under bark of dead tree trunks, mosses and lichens, on living tree trunks and under stones/leaf litter) with the collection of six replicates/microhabitat/site (total 18 samples per site).
For each spider species collected, we recorded information on sampling technique used, site location code (Table 1), number of captured specimens and gender whenever possible. Information on species distribution follows the World Spider Catalog (Platnick 2013), updated with records from regional checklists of the Macaronesian archipelagoes (Cardoso and Crespo 2008, Macias 2009). Furthermore, distribution maps are provided for each endemic species found in the Laurisilva of Madeira island (Figs 1,2,3,4,5). The specimens collected are deposited in the entomological collection of the Faculty of Sciences of the University of Lisbon (Portugal).

PTMAD0001, PTMAD0002
Description: The study took place in Madeira island located in the Atlantic Ocean, approximately 1000 km from the Iberian Peninsula. Sampling was conducted in thirty-six sites (Table 1), mostly in Laurisilva patches, the only exceptions being two sites on an alpine meadow near Pico do Areeiro, one of the highest mountains on the island. Laurisilva sites were mostly located in pristine or near pristine forest patches (Figs 1a,6) where the dominant tree species included some of the following Madeiran or Macaronesian endemics: Clethra arborea, Erica platycodon maderincola, Laurus novocanariensis, Ocotea foetens and Persea indica. Additional information on plant species composition and conservation status of sampling sites can be found in Neves et al. 1996. Sampling sites and distribution of endemic spider species in Madeira island. See Table 1 for detailed information on site location. Circles: present records; rhombuses: records from Schenkel (1938); squares: records from Denis (1962); triangles: records from Wunderlich (1987). Distribution of endemic spider species in Madeira island. Circles: present records; rhombuses: records from Schenkel (1938); squares: records from Denis (1962); triangles: records from Wunderlich (1987).  Distribution of endemic spider species in Madeira island. Circles: present records; rhombuses: records from Schenkel (1938); squares: records from Denis (1962); triangles: records from Wunderlich (1987); stars: records from Wunderlich (1992). Distribution of endemic spider species in Madeira island. Circles: present records; rhombuses: records from Schenkel (1938); squares: records from Denis (1962); triangles: records from Wunderlich (1987); stars: records from Wunderlich (1992).

Distribution: Canary Islands, Deserta Grande island, Madeira island
Notes: This species was recently found for the first time in Madeira archipelago in an erosion cave of Deserta Grande island (Crespo et al. 2013), and its distribution has been now extended by this study to Madeira island.  (Fig. 1c).

Clubiona decora
Notes: This is the first time this species was captured after its description (Schenkel 1938). The previous record only included adult specimens from Caramujo, but this study shows that the species has a wider distribution including many Laurisilva sites in Madeira island. Notes: This is a new species to science, to be described in a future work. As the previous species of Ceratinopsis, it also seems to be restricted to Laurisilva. Notes: This is the first record of this endemic species since its original description, which is based on females only and fails to include the precise location of the type material. The specimens collected in this study will be used in a future work to describe the male of this endemic species.  (Fig. 3a) Notes: This endemic species seems to be restricted to Laurisilva. The two records outside Laurisilva (Denis 1962) are most likely artifacts due to imprecise location of sampling localities and/or reductions in Laurisilva cover during the last decades.  (Fig. 3b) Notes: Distribution data on L. impudicus is scarce consisting of only 4 records. This species seems to be associated with Laurisilva and the lack of congruence between historical data (Denis 1962) and the current distribution of Laurisilva may be explained by the lack of precision in the identification of sampling sites and/or by the decrease of Laurisilva cover in the last decades.  (Fig. 3d) Notes: This is the third record of this species. Turinyphia maderiana seems to be restricted to Laurisilva.

Discussion
Fifty two spider species and 3145 specimens are accounted in this study. Out of these, 17 are Madeira's endemics species (SIE) mostly restricted to Laurisilva. Range for most endemic spider species was expanded by this study (Figs 1, 2, 3, 4, 5) and six endemics [ Hahnia insulana Schenkel, 1938, Frontinellina dearmata (Kulczynski, 1899, Lepthyphantes mauli Wunderlich, 1992, Trogloneta madeirensis Wunderlich, 1987, Metastridulans Wunderlich, 1987 and Dipoenata longitarsis (Denis, 1962)], were collected for the first time after their original description. A major finding of this study was the discovery of two species new to science, a species of Ceratinopsis Emerton, 1882, and a species of Theridion Walckenaer, 1805, apparently restricted to Laurisilva. The remaining thirty five spider species were assigned to one of the two following categories: nonendemic native (for both macaronesian endemics and species with a wider distribution) and introduced species. The inclusion of each species in these categories was based on distribution and species ecological traits. Sixteen species were classified as introduced species, including Poeciloneta variegata (Blackwall, 1841) and Tetragnatha intermedia Kulczynski, 1891, two new records for Madeira island. Two species with uncertain identity, Macaroeris sp. and "Linyphiidae unidentified species" were considered as non-endemic native species.

Species richness considerations
The highest values of species richness were found in sites 18, 7, 11, 12, 25 and 33, which were also among those with the largest number of endemics (Fig. 7, Table 2). The lowest species richness was recorded in site 31, with only 1 native species. Abundance was maximum on alpine meadow sites (15, 16), mostly due to the presence of several abundant species like Zodarion styliferum, Haplodrassus dalmatensis and Zelotes civicus. These ground-dwelling species are quite vagile and are among the top invertebrate predators of Madeiran altitudinal meadows. In most study areas, introduced species were more abundant than natives, with some exceptions, such as sites 11, 12, 13, 18, 25 and 34, where higher proportions of specimens from endemic species were recorded (Fig. 8)  This is the first time that a large set of standardized samples of spiders from Madeira island is reported. The distribution of several Madeiran endemic species was greatly increased, particularly Macarophaeus cultior, Hahnia insulana, Turinyphia maderiana, Trogloneta madeirensis, Rugathodes madeirensis and Episinus maderianus. Other endemics were recorded in fewer sites, but in most cases such species are still associated with Laurisilva. In some instances, the historical distribution of endemic spider species includes areas no longer covered by Laurisilva. An example is a record of Centromerus variegatus from Santo da Serra (Denis 1962), a site now covered by exotic forest but which was presumably dominated by endemic Erica shrubs just some decades ago. Species richness values per site, highlighting the relative proportion of endemic, native and introduced species. The data of each site combines information from both pitfall and direct sampling. Suppl. material 1.

Figure 8.
Scatter plot of the log-transformed species richness of endemic and native species versus exotics for each sampling site. Red circles -from both pitfall trapping and direct sampling; Green squarespitfall data only. Suppl. material 1.
The need to apply a combination of sampling techniques in a standardized design in order to assess species composition and structure of ecological communities has been strongly emphasized in the last few years by different authors , Borges et al. (2005), Borges et al. (2006), ). Accordingly, combining two complementary sampling techniques was determinant in our study to find a variety of spider species that use the habitat in different ways.
In complex habitats like Laurisilva, the prospection of microhabitats such as fallen logs and trees, tree holes, under tree bark and epiphytes allows the finding of species with highly specialized niches. Furthermore, several spider species seldom move or do not move at all at ground level, making direct sampling the most suitable detection technique. From the 52 spider species reported, 16 were collected exclusively using pitfall traps while 22 other species were only found by direct sampling. On the other hand, direct sampling proved to be an effective method to sample Madeiran endemics since most of the sampled endemic spider species were only collected using this technique (see also Fig. 8).
The original descriptions of some Madeiran endemics do not allow inequivocal identification since they were based on very few specimens or on specimens of a single gender. Thus, taxonomic revisions of several genera or families (i.e. Dysdera Latreille, 1804, Hahnia C.L. Koch, 1841, or the Linyphiidae) are mandatory for a better understanding of Madeiran spider biodiversity. Besides the merits of the present study in improving the knowledge on the distribution of spider species in Madeira island, it also allowed the collection of a number of specimens that will be the aim of future taxonomic studies, such as species redescriptions and formal descriptions of new species to science.
Madeira Laurisilva presents a unique spider biodiversity and efforts should be made towards improving knowledge on the abundance and distribution of endemic species as well as to develop a strategy for their conservation and monitoring.