Distinctive Collembola communities in the Mesovoid Shallow Substratum: First data for the Sierra de Guadarrama National Park (Central Spain) and a description of two new species of Orchesella (Entomobryidae)

Two new species of the genus Orchesella Templeton, 1836 have been identified following intensive sampling in the Colluvial Milieu Souterrain Superficiel (Mesovoid Shallow Substratum, or MSS) of the Sierra de Guadarrama using Subterranean Sampling Devices (SSD). The data were obtained from the first extraction of the traps between May and October of 2015. During a study of the Collembola taxon, 32 different genera (61 species) were identified. The highest representative genus presence in almost all traps was Orchesella, with two new species. One of the two species described had been misidentified until this study was carried out, indicating that their preferential habitat had not been sampled; the second species had never been identified. The community of the Orchesella species in the Colluvial MSS was investigated, leading to the conclusion that this environment has its own assemblage of characteristic species. The opportunity to study specimens that belong to five species of the genus Orchesella, including three previously recollected, has allowed for obtaining reliable information regarding their macrochaetotaxy. A part of this chaetotaxy is proposed as a useful diagnostic tool for the species of the genus. In conclusion, it can be affirmed that this study has demonstrated that the Colluvial Mesovoid Shallow Substratum (Colluvial MSS) has its own fauna, and it supports the hypothesis that it constitutes a new biotope, at least for Collembola.


Introduction
The study of subterranean habitats has generally focused on caves, pits and even mines (of anthropogenic origin) [1,2] although shallow subterranean environments [3] and specifically PLOS  favour the formation of scree-talus, and therefore these substrates have been excluded from underground surveys (Fig 2). The park consists of three bioclimatic floors, which respond to the typical model of a mountainous Mediterranean enclave: the supramediterranean, oromediterranean and crioromediterranean floors. On the supramediterranean floor (1300-1700 m a. s. l.), the melojo oak (Quercus pyrenaica Willd.) is the most common species, although it has been depleted in favour of the Scots' pine (Pinus sylvestris L.), which mostly grows in the upper supramediterranean floor but is also found in the oromediterranean and the crioromediterranean floors of some south-facing slopes [43]. These pine forests, whether natural or naturalised, comprise the second most extensive vegetation unit of the park, forming dense forest masses with little fragmentation [36] (Fig 2). The oromediterranean floor (1700-2200 m a. s. l.) is characterised by the presence of Juniperus communis alpina (Suter) Celak. (creeping juniper) and shrubs of Cytisus oromediterraneus Rivas-Martinez et al. ('piorno serrano'), although these species can enter the supramediterranean floor, reaching lower levels than is optimum. On the oromediterranean floor, these species include shrubs of timberlines and Adenocarpus hispanicus (Lam.) DC. or Erica arborea L. This vegetation cover alternates between grasslands, rocky outcrops and scree-talus. It is the most extensive vegetation in the Sierra de Guadarrama National Park [36], which forms the characteristic habitat of the summits of this mountain along with the areas covered by scree-talus (Fig 2). Finally, the crioromediterranean floor (2200-2400 m a. s. l.) has a combination of psychoxerophilic grasses and Festuca curvifolia Lag. ex Lange, which plays an important role in soil protection in extremely harsh climatic conditions due to its coverage [44]. In addition, there is an entire cohort of colluvial specialist vegetation [36] as well as large colluvial and morrenic deposits that extend along the oromediterranean floor.
In the Sierra de Guadarrama, the climate is a cold continental Mediterranean with dry and fresh summers. Its marked continentality due to its remoteness from the influence of the Mediterranean Sea and the Atlantic Ocean is not incompatible with the presence of a remarkable variety of microclimates. The microclimatic differences result from the special arrangement of the three mountain ranges, which condition the circulation of the air masses [45]. In addition, the topographic and geographical diversity (orientations, slopes, altitude) of each of these mountains decisively influence meteorological parameters, such as temperature, insolation, precipitation and the accumulation and persistence of snow [46]. The cold period extends beyond winter for about six months, while the dry period does not last longer than one month. The remaining five months are characterised by intense biological activity due to a combination of temperature and humidity that promotes the vegetative period of the plants [47]. Meteorological data of the port of Navacerrada (1894 m a. s. l.) recorded over a 20-year period yielded the following average annual values [48] that can guide the tendency of the main climatological variables of the Sierra de Guadarrama: maximum temperature 10.7˚C; minimum temperature 3.1˚C; 142 days of frost; and 1223 mm of precipitation over 111.3 days, of which 78 days was snow.

Sampling
The sampling was performed mainly using Subterranean Sampling Devices (SSD) similar to those used in other underground surveys in Iberian enclaves [12,25,23,49,50,51,14]. The SSDs consisted of a PVC cylinder 11 cm in diameter and 1 m in length, which had numerous perforations that were 8 mm in diameter in its lower half (Fig 2F). Once buried vertically in a suitable substrate, the top was covered with the ground surface (Fig 2C-2E). A pitfall trap was slid inside using a nylon thread that precisely matched the tube wide ( Fig 2G). Previously, the trap was partially filled with a preservative liquid (1,2-propanediol) and fed with a vial containing strongly odorous cheese (Fig 2G and 2I). Inside some SSDs (SSD-5, SSD-7, SSD-12, SSD-14, SSD-17 and SSD-23), a digital thermometer programmed to record temperature at one-hour intervals (Fig 2G) was installed and was protected by a plastic bell. In the La Mujer Muerta range, SSDs of 0.5 m perforated along their entire length except for the upper 20 cm were installed ( Fig 2H) along with the four 1 m SSDs (SSD-1, SSD-2, SSD-3 and SSD-4) ( Fig  2F). The purpose of the SSD model of 0.5 m was to collect the fauna of the most superficial MSS and thus allow for comparing the fauna collected with that extracted from the SSD of 1 m. Two pitfall traps (TSPs), which contained the same type of preservative and bait as the SSDs, were buried in a visible Colluvial MSS in a forest talus of the Puerto de Navafría. To achieve this, two horizontal holes of about 80 cm in length were created ( Table 1). The coordinates used are Universal Transverse Mercator coordinate system (UTM).
The enclaves sampled were located within the national park (32 localities), and to a lesser extent, within the area declared Peripheral Protection Area (2 localities) ( Table 1 The biological samples (specimens collected from May to October 2015) were transported to the laboratory and were preserved in 70% ethanol until the taxonomic study began.

Mounting and observation
Some specimens were cleared in Nesbitt's fluid, and after washing for one hour in 70% Ethyl alcohol, they were mounted in Hoyer's medium for the compound microscope observation in a phase contrast and DIC. Some specimens were decapitated, immersed in a saturated solution of KOH and emptied before being mounted on the slides. Some specimens were examined using a Scanning Electron Microscope (SEM). Specimens from 70% ethyl alcohol were slowly rehydrated in a decreasing series of concentrations (60%, 50%, 40%, 30%, 20%, 10%, distilled water) over 24 hours. Once in distilled water, they were fixed in 4% glutaraldehyde in a cacodylate buffer for 24 h and then transferred to sucrose 0.25 M for 24 h. The specimens were then dehydrated up to 100% ethyl alcohol. Complete desiccation was achieved using the CO 2 critical point technique. Samples were then covered with a 16-nm thin layer of molecular gold using an Emitech K550 sputter coater. Observations were performed using a Zeiss Digital Scanning Microscope 940 A.

Taxonomy
The combined use of color and morphological characters (macrochaetotaxy included) allowed for the identification of the species and provided useful descriptions for species identification. Potapov and Kremenitsa used coloration and macrochaetotaxy (some selected areas), the antennae length, width and shape, the eye structure, leg length, claw complex and manubrial organ (males) [33]. The array (or group) of characteristics proposed by Jordana and Baquero [32] and the chaetotaxy nomenclature of [31], which is based on a constant and generally visible set of morphological characters [52,53,54], has proven useful for the identification of species that belong to Entomobryini. It is proposed that some of the areas that conform to the simplified formula of [32] for Entomobrya could also apply to Orchesella, and an area for the head macrochaetotaxy (H6) has been added, which indicates the presence of chaetae S 0 and S 2 and their accessories.

Fauna associated with the Colluvial MSS
For the identification of the Collembola species examined in the present study, a total of 42,240 specimens (20% belonging to the genus Orchesella) were studied individually. This work was conducted at the Department of Environmental Biology, University of Navarra, after the triage was completed at the University of Alcalá for over 61,038 specimens. This means that 69 per cent of the specimens caught in the total traps belonged to the Collembola taxon. Throughout the first collection of sampling traps from the MSS of the Sierra de Guadarrama, only for the fauna of Collembola, a total of 32 genera and 61 species were identified. Of these 61 species, 18 are likely new to science. The taxon Entomobryidae, with five genera and 18 species, was the most uniformly distributed throughout the entire study area. The genus Orchesella, represented by two new species, was present in almost all traps (absent only in four). Thus, these two species are considered the most representative of the Colluvial MSS of the Sierra de Guadarrama.  ETYMOLOGY.-The species name follows the ecological niche. This was the first time it had been studied in Sierra de Guadarrama, where it was found.
Ground color of the body pale; eyes black; head with light blue pigment on lateral and posterior head; antennae slightly blue at the end of the second part of the first segment, dark blue in the first part of the second segment and slightly blue for the last two segments; Thorax (Th) II-Abdomen (Abd) II with lateral, dorsolateral and dorsal-central longitudinal dark blue stripes, broadened at the end of each segment; Abd III with a transversal dark blue irregular band; Abd IV characterised by a continuation of the pigment patches of the forwarded segments; Abd V with dorsal and lateral spots; Abd VI without pigment; legs: femur with terminal lateral blue pigment and tibiotarsus homogenously light blue. Juveniles with a similar coloration.
Head. Ratio antennae/head = 4.04 (n = 14). First and second sub-segments of Antennal segment (Ant) I thick, twice as Ant II. Distal part of Ant III and Ant IV annulated and verticillate (Figs 4A and 4B and 6C and 6D), without apical bulb; Ant III sensory organ as in Fig 4B  (detail). Labral papillae rounded, with pointed and curved tip, as a hook. Prelabral and labral chaetae all smooth (4/554). Lateral process (or differentiated chaeta) of outer labial palp (E) not reaching the middle of its papilla. Anterior labial row with four smooth and subequal chaetae; posterior labial row internal to chaeta E with 6-8 chaetae per side, all ciliated; chaetae E, L 1 and L 2 ciliated. Outer lobe of maxillary palp with four smooth chaetae in addition to the two basal chaetae. 8 + 8 ommatidia, G and H smaller (Fig 6A).
Macro-chaetotaxy (Figs 7A, 7E and 8A, Table 2). Head: H1 area with five to seven An chaetae, some as mc (mesochaetae); H2 area with 2-3 Mc (Macrochaetae) and 2-3 mc; H3 area with S' 0 present; H4 area with two Mc; the S series (from S 4i to the eye) with 10-13 chaetae, 3-4 as Mc (chaetae of different sizes); H5 area with two Mc; H6 area with four or five Mc (sometimes as mc). Th II: T1 area with 10-12 Mc (n = 4); T2 area with 7-10 Mc, m 5 not seen in one of the four specimens studied. Th III with three Mc above pseudoporus, and 12-13 chaetae in total; lateral sensillae not identified. Abd I with five Mc around the pseudoporus (a 1 , a 2 , m 2i , m 2 and m 3 ), and four towards lateral (a 3 , m 4i , m 4 and a 6 ); lateral sensillae not identified. Abd II with four Mc in A1 area (a 1 , a 2i , a 2 and a 3 ), and four on A2 area (m 3 , m 3ei , m 3e and m 3e2 ); sensilla 'as' above m 3 (between m 3 and a 3 ); m 2 and a 5 as bothriotricha. Abd III with one Mc on A3 area (a 1 ), three on A4 area (a 2 , a 2a and a 3 ) and two on A5 area (m 3 and m 3e ); sensilla 'as' above m 2 (between m 3 and a 3 ); a 5 , m 2 and a 5 as bothriotricha, three Mc lateral to external bothriotricha. Abd IV: A6 area without Mc; A7 area with three Mc (A 1 , B 1 and B e2 ); A8 area with three Mc (A 3 , B 2 and B 3 ) and four mc (A 03 , A e3 , C 1 and E 1 ); A9 area-than included the pseudopore, in a more advanced position than in other species of the family that were closer to  Biology. The study of the intestinal content of the specimens mounted on the slides allowed for determining that they feed mainly on amorphous organic matter. In addition, in almost every specimen, fungal hyphae and spores were observed. There was also a noticeable quantity of crystals (presumably small sand grains).
Remarks. The coloration of the species O. mesovoides sp. nov was completely different from the seven species mentioned in the Iberian Peninsula, though it was somewhat similar to the original description of O. eolia (from Stromboli Island, cited from the Sierra de Guadarrama) [55]. Two of these specimens cited by Selga were observed. The first was from Collado de la Mina (or Lijar) in Alto del León (mine of wolframio) (MNCN_Ent 144262). Another specimen from Sierra de Navacerrada was a juvenile that was cleared and mounted sideways and the chaetotaxy is not visible, (MNCN_Ent 144270). The first belonged to the O. mesovoides sp. nov. due to its coloration with a less extended pigment, and the macro-chaetotaxy was also similar.
It was possible to identify differences in the pigment patterns between the new species and O. eolia, especially in the Th II and Abd II-V segments. The ratio Abd IV/III in O. eolia (1.48) was lower than that for the species O. mesovoides sp. nov.  ETYMOLOGY.-The species name follows the ecological niche of where it was found (rock detritus and soil accumulated at the foot of a mountain) from Latin colluvium.
Ground color of the body uniformly pale; eyes black; head with light blue pigment only lateral to eyes and insertion of Ant I; antennae pigmented on dorsal Ant I b, Ant II a, dorsal Ant II b, and last two segments; only the lateral Th III, Abd II-III, posterior lateral Abd IV, lateral Abd V, and in some specimens, external distal trochanter, femur and tibiotarus, had some pigment.
Head. Ratio antennae/head = 4.30 (n = 11). First and second sub-segments of Ant I thickness, twice that of Ant II. Distal part of Ant III and Ant IV annulated and verticillate (Fig 4C  and 4D) without apical bulb; Ant III sensory organ similar to that of O. mesovoides sp. nov. Labral papillae rounded, with pointed and curved tip, as a hook; in some specimens hook asymmetrically bifurcate, sometimes with a bristle, with one to three ramus. Prelabral and labral chaetae all smooth (4/554). Lateral process (or differentiated chaeta) of outer labial palp (E) not reaching the middle of its papilla. Anterior labial row with four smooth and subequal chaetae; posterior labial row internal to chaeta E with 13-15 chaetae per side, all ciliated; chaetae E, L 1 and L 2 ciliated (E and L 1 smooth in some specimens). 8 + 8 ommatidia, G and H smaller; the eye H, seen at SEM, poorly developed externally (Fig 6B). Outer lobe of maxillary palp with four smooth chaetae in addition to the two basal chaetae (Fig 5J).
Furcula. Males with manubrial organs with 3-4 small chaetae grouped similar to a brush. The smooth area of distal dens 2 times the mucronal length; basal spine on mucro reaching or slightly surpassing the basal tooth; teeth of similar size (Fig 5I).
Biology. The study of the intestinal content of the specimens mounted on slides allowed for observing amorphous organic matter, although some pollen grains were also present in some specimens that likely belonged to Pinus sylvestris, which are common in the area.
Remarks. The coloration, which was an almost completely yellowish-white, of this species was similar to O. flavescens, which has two dorsolateral longitudinal lines from the head (behind the eyes) to the segment Abd VI in most specimens. It was possible to deduce part of the formula proposed in this work for O. flavescens using the information provided by Stach [28]:? -?-?-?-?/3-6/8-4/1-3-3/3-7(8)-3-2-0 (?, unknown), which makes it possible to differentiate it from O. colluvialis sp. nov. (Table 2) (other comparison among the different species of Orchesella cited in this paper is presented in Table 3). Other species recorded in Europe have some specimens from some populations with very little pigmentation, but they are always cited together with specimens with apparent pigmentation. In the case of O. colluvialis sp. nov., which was captured in hundreds of specimens in the sampling of this study, all specimens had pigmentation exclusively on the sides of the Abd VI (as shown in Fig 3D,  Description. Body lengths of the specimens were 3.10-5.20 mm (mean 4.30 mm, n = 4) (Fig 3E). Head. Ratio antennae/head = 3.05 (n = 2). Coloration is as described for the species [56]. First and second sub-segments of Ant I thick, some less than twice that of Ant II. Distal part of Ant III and Ant IV not annulated or verticillate, perhaps only a hint of annulation in the most distal part (Fig 4E) without an apical bulb. Labral papillae rounded, with pointed and curved tip, as a hook. Prelabral and labral chaetae all smooth (4/554). Lateral process (or differentiated chaeta) of outer labial palp (E) not reaching the middle of its papilla. Anterior labial row with four smooth and subequal chaetae; posterior labial row internal to chaeta E with 11-13 chaetae per side, all ciliated; chaetae E, L 1 and L 2 ciliated. 8+8 ommatidia, G and H smaller.
Furcula. Males with manubrial organ with some chaetae grouped similar to a brush. The smooth area of distal dens was 1-2 times the mucronal length; basal spine on mucro reaching or slightly surpassing the basal tooth; teeth of similar size (Fig 5L).
Remarks. This species has a wide distribution in the Northern Hemisphere. The macrochaetotaxy contributed to this paper. Description. The body lengths of the specimens studied were 3.65-4.40 mm (mean 4.00 mm, n = 7) (Fig 3F).
Head. Ratio antennae/head = 3.18 (n = 3). Coloration is as described for the species [56]. First and second sub-segments of Ant I wider than Ant II. Distal part of Ant III and Ant IV not annulated or verticillate, perhaps only a hint of annulation in the most distal part (Fig 4F  Table 2) not shared between the compared species; left and down, percentage of dissimilarity between species. https://doi.org/10.1371/journal.pone.0189205.t003 Two new Orchesella distinctive in the MSS from Sierra de Guadarrama National Park and 4G), without an apical bulb. Labral papillae rounded, with pointed and curved tip, as a hook. Prelabral and labral chaetae all smooth (4/554). Lateral process (or differentiated chaeta) of outer labial palp (E) not reaching the middle of its papilla. Anterior labial row with four smooth and subequal chaetae; posterior labial row internal to chaeta E with 7-9 chaetae per side, all ciliated; chaetae E and L 1 ciliated, L 2 smooth. 8 + 8 ommatidia, G and H smaller. Body. Abd IV/III ratio 1.30-2.11 (n = 7). Trochanteral organ with 80-90 special chaetae (Fig 5D). Tibiotarsus without sub-segmentation, although in some specimens, it was bent slightly beyond its midpoint; spiniform chaetae with appressed ciliation (apparently wide, smooth chaetae): 5-10 on leg 1; 10-12 on legs 2 and 3. Claw with one outer tooth, slightly more basal than the level of a pair of stout lateral teeth, and four inner teeth (two paired at 40%, a first unpaired at 70% and a last unpaired, with a significant development at 90% of the internal edge total length). Empodial appendage lanceolated and concave, with an outer tooth inserted at 50% of its total empodial length; empodial appendage/claw length ratio 0.67 (0.64-0.70, n = 4); smooth chaetae on tibiotarsus similar in length to empodial appendage (0.97-1.03, n = 4); tenent hair slightly longer (1,27 times) in length than empodial appendage (n = 1) (Fig 5M). Anterior face of ventral tube with many chaetae (>50), including 2+2 Mc on distal area; posterior face with >50 chaetae, including 1+1 Mc on middle-posterolateral area, and 2 +2 ms on basal area; apical flaps with >50 smooth and ciliate chaetae, some of them bigger than other.
Furcula. Males with manubrial organs with four chaetae grouped similar to a brush. The smooth area of distal dens 3-4 times the mucronal length (although in one of the studied specimens, it was only nearly 2); basal spine on mucro reaching or slightly surpassing the basal tooth; distal tooth bigger than basal (Fig 5N).
Remarks. The specimens studied were identified by color. In this study, the information regarding the macro-chaetotaxy of this species is presented for the first time, which will allow for comparisons when studying the specimens identified by other researchers either in the past or in the future. A wide distribution may be attributed to the species based only on coloration changes when the macro-chaetotaxy is used.

Ecology
The study allowed for obtaining data regarding the microclimatic conditions of the sampled habitat, mainly of temperature, during the sampling period from June to September of 2016. It was observed that the highest monthly average temperatures were recorded in July, with 17˚C in the highest altitudes and 23˚C in the lowest altitudes, which decreased below 15˚C in late August. The temperatures in the MSS never exceeded 26˚C and did not decrease below 5˚C during the cited sampling period. In addition, a temperature range between 5 and 18˚C was observed in the traps located in the surroundings of 2000 m. a. s. l., and in the surroundings of 1500 m. a. s. l., the temperature range was between 10 and 25˚C (Fig 11). The presence of the two species along the Sierra de Guadarrama was fairly uniform ( for higher altitudes (Fig 12 and inset A). Considering the average number of animals per trap (to weigh the quantity of traps per orientation), it was observed that the vector sum of the abundances for the species O. colluvialis sp. nov. was located in the W-NW orientation, whereas for the species O. mesovoides sp. nov., it was N-NE (Fig 12, inset B). The presence of pollen granules in the species O. colluvialis sp. nov. may be related to its more abundant presence in lower areas of the mountain range where the tree vegetation, Pinus sylvestris, is more abundant.
The two new species described have similarities with respect to their distribution and abundance at different places, but it can be argued that the species O. colluvialis sp. nov. seems to be more adapted to the underground environment, which could be based on several factors: its coloration with little pigment, a greater reduction of the eyes (the G and H eyes was far more reduced in the optical microscope, which was confirmed when observed at SEM) ( Fig 6B) and more S-chaetae. O. mesovoides sp. nov. had been referenced from Sierra de Guadarrama (erroneously identified as O. eolia), and given the lack of previous samplings in the subterranean environment, these references suggest that this species has less dependence on the subterranean habitat. Nevertheless, it has never been recorded at the surface. The abundance by trap varied between a few specimens and at a maximum of 700 (in this case, the species O. colluvialis sp. nov.). For traps with two-level catches, no difference was observed for the two species described. In total, 1956 specimens of O. mesovoides sp. nov. and 5929 specimens of O. colluvialis sp. nov. were caught. The data confirmed that the species are abundant in the massif. They have not been captured previously, with the exception of the capture of the specimens identified as O. eolia by Selga in Collado Alto and Valsaín, which demonstrates their absence on the surface in this area. Based on the findings it can be affirmed that the Orchesella communities of the Colluvial MSS of Sierra de Guadarrama have their own species. Interestingly, other Orchesella species appeared to inhabit the Colluvial MSS in other regions of Europe [16,17,22] and even in the Iberian Peninsula [12,14]. Thus, future works should reveal a significant hidden diversity of species of the genus as well as different species assemblages in the MSSs of different regions.
In addition to the work of the researchers cited above and to Bonet and Acon [60,61], who collected Collembola from 1929 until recently, and with the exception of two specimens identified as O. eolia found in a gallery of a mine, there have not been any additional registers of the Orchesella species described in this paper. Nevertheless, they have been proven to be abundant in the colluvial MSS.
It was possible to compare the two new species described in this paper with the species O. villosa, O. cincta and O. eolia ( Table 3). The macrochaetotaxy of O. flavescens sensu Stach [28] has been studied for comparisons with O. colluvialis. It was not possible to study the species O. quinquefasciata. While comparing the macrochetotaxy and other morphological characters, it was found that with these characteristics and without color, these species can be differentiated. This is of enormous importance because until now, the identification of the genus Orchesella was based almost only on the coloration. Moreover, the sensilar chaetotaxy (S-chaetae and Smicrochaetae) was examined in detail, since it has been used for other genera [62], although it is difficult to observe in species with large specimens. developing this project: Gonzalo Pérez-Suárez, Alberto Jiménez-Valverde, Alberto Sendra, Pablo Barranco, Alberto Tinaut, Luis Subías and Juan José Hererro-Borgoñón. Finally, it should be recognised that the field work could not have been completed without the valuable collaboration of the authorities responsible for the national park as well as other workers (biologists, forestry agents and staff of the Oficina de Atención a la Población Local-Sector Lozoya-and the Centro de Visitantes-Peñalara). In particular, we extend our thanks to Patricia Riquelme, Pablo Sanjuanbenito, Juan A. Vielva, Javier Donés, Marisol Redondo, Ignacio Granados, Á ngel Rubio, César Martín, José Carrillo, Miguel Á ngel Palomar, Á ngel Velasco, Germán Mato, Manuel Criado, Enrique Calvo, Federico Madejón and Montserrat Sanz. The anonymous reviewers provided excellent comments and suggestions that improved this manuscript.