North PyreNeaN PoPulatioNs of Megabunus diadeMa ( fabricius , 1779 ) ( arachNida : oPilioNes ) are characterized by highly male-biased sex ratios

Megabunus diadema (Fabricius, 1779) is an Atlantic and European harvestman species characterized by a discontinuous distribution from Scandinavia to the Iberian Peninsula. With very few male individuals ever observed in the field until now, the biological uniqueness of the species lies in its reproduction mode, hitherto regarded as asexual, facultative parthenogenesis. Based on a large sample of 741 sexed individuals, the study indicates a sex ratio much higher than what was formerly known, equal to 65.58% of males. Locally varying from 0 to 100% (median 75.5% of males), the sex ratio depends indeed on the altitude and the phenological cycle: the proportion of males decreases with increasing altitude and increases gradually during the spring to reach a plateau in summer. By describing populations locally dominated by male individuals and providing new information on the spatial and temporal patterns of tertiary sex ratio, we question the currently admitted reproduction mode of the species which could be normally sexual, at least locally, rather than asexual. A distribution map of the species on the northern slope of the Pyrenees is provided for the first time. Our study also complements the distribution for the southern slopes of the Pyrenees and the rest of the Iberian Peninsula published recently by Merino-Sáinz et al. (2013).

If the outlines of its overall biogeographic area are better known, there are still some gaps about the ecological envelope of the species, especially at the meso-and microhabitat scales.For example, the limits of its altitudinal distribution are still imprecise and ecological versatility is yet to be characterized.The biggest gaps and contradictions concern the reproduction of the species.Many authors have already pointed out that Megabunus diadema is a facultative parthenogenetic species (Phillipson, 1959;Feliú, 1981;Stol, 2005).The majority of the information provided in the literature as well as, marginally, in discussion forums (webography not supplied) report mostly fe-male individuals sightings.For example, in England, Phillipson (1959) found a single male individual versus 406 females during his survey.In Norway, the proportion provided by Stol (2005) is comparable, 3 males to 261 females.In Spain, the tendency to over-representation of females is also reported by Merino-Sáinz et al. (2013), although the proportion of males is slightly higher than elsewhere (8 males to 51 females) but they recognize possible local variation of this proportion, namely in the Huesca region where the largest number of males was found (7 males off the total of eight found).By contrast, D'Amico & Besson (1995) already described an overrepresentation of males in the population monitored in the Ossau Valley, on the northern slope of the Pyrenees.
Based on current knowledge, Megabunus diadema appears to be a species with highly biased sex ratio in favour of females throughout its range, with the exception of the Pyrenean populations.The main objective of this work is to quantify for the first time this variation in the sex ratio of the north-Pyrenean populations by examining the patterns of variation in spatial dimensions (focusing here on altitude) and temporal dimension (here adult development period).
This work is also used to update the distribution of the species on the northern slope of the Pyrenees and to specify the limits of its altitudinal distribution.

methods
Data were collected between 1988 and 2014 (see also D 'Amico, 1988;D'Amico & Besson 1995) by sight sampling, methodically searching for individuals of the species on favorable substrates (rocks, trunks ...).These original presence data (see Annex), combined with historical data (Simon, 1881), and other published data (Evenou, undated;Ledoux & Emerit, 2006), allow to establish the distribution map of the species on the northern side of the Pyrenees.Data published by Merino-Sáinz et al. (2013) for the south side are not taken into account.
The characterization of the sex ratio and its spatial and temporal variation was made from a subset of this data set, putting as strong conditions a reliable identification of males and females and a minimum number of 3 individuals in the sample.In accordance with the definitions proposed by West (2009), the sex ratio is the proportion of males in a population.Although the term sex ratio is used indiscriminately throughout this paper, figures provided in this study relate only to the tertiary sex ratio, defined as the ratio for mature organisms (adults).Primary sex ratio (ratio at the time of conception) and secondary sex ratio (ratio at birth) are not evaluated in this work.
The patterns of spatial and temporal variation in the sex ratio are analysed in a descriptive manner using multiple boxplots built with the commands "geom_boxplot» and «geom_jitter" of ggplot2 package for the free software R (R Core Team, 2012).These commands of the package ggplot2 (Wickham, 2009) allow to superimpose all raw data as points to the conventional box plots.In boxes presented here, the horizontal line represents the median and the upper and lower "hinges" correspond to the first and third quartiles (the 25th and 75th percentiles).

geographic distribution
The species has a wide distribution over the whole French Pyrenees (Figure 1), with an apparent under-representation in the easternmost part.Indeed, we provide the first records of this species in the departments of Haute-Garonne and Ariège.To date, no record of the species has been made in the department of Aude.The species was already cited from the Pyrénées-Atlantiques (Simon, 1879;D'Amico, 1988) and Hautes-Pyrénées (Simon, 1881) whilst its presence was first recorded recently (Ledoux & Emerit, 2006) in the Pyrénées-Orientales from two close locations.

altitudinal distribution
The altitudinal distribution profile for Megabunus diadema in the Pyrenees is established from an original dataset of 82 records.The species is found between 220 and 1920 m, with a mean altitudinal value of 1143 m (median value 1182 m).

sex-ratio
The characteristics of the sex ratio are based on a sample of 54 records, with an average of 13.7 individuals per record (range: 3-37 individuals).In total, 741 individuals were sexed during the study, 486 males and 255 females, corresponding to a total value of sex ratio equal to 65.58%.In statistical terms, for the entire dataset, the sexratio varies between 0 and 100%, with an average of 64.0% and a median of 75.5%.
The great variability in the sex ratio results in a decrease in the proportion of males as altitude increases (Figure 2): high (ca 80%) between 400 and 1200 m, it recedes quickly in its upper altitudinal range to a median value of around 30%.
The sex ratio shows a well marked pattern in temporal variation, indicated by a gradual increase during the spring and a plateau in the summer (Figure 3).Lower than 50% in May, when adults appear, it reaches a median value of around 90% in July and August.

discussion
A distribution map of the species on the northern slope of the Pyrenees is provided for the first time, with mostly new records from all the french Pyrenees.Apparent under-representation in the easternmost part (namely departement of Ariège and Aude) is probably an artifact, resulting from a relative lower field survey pressure in this area.Our study complements the one recently compiled by Merino-Sáinz et al. (2013) for the southern side of the Pyrenees and the rest of the Iberian Peninsula.Our work provides new and original information on the variability of the tertiary sex ratio of this species whose notable originality lies in its reproduction mode: considered until now as facultative parthenogenesis, with very few male individuals observed in the nature, our study (involving 741 individuals in total, the largest number of individuals ever studied in this species) leads to an opposite conclusion, at least in the northern Pyrenees, and emphasizes for the first time the possibility for this species to be a sexually reproducing one.

distribution
In his review on mountain harvestmen, Beron (2002) indicates that many species can live or survive at high altitude.In Europe, he cites at least 47 species reaching or exceeding 2200 m; they belong to 6 families (Phalangidae, Sclerosomatidae, Ischyropsalididae, Trogulidae, Sabaconidae and Nemastomatidae).Of these, only 14 species are observed at 2500 m and only 3 reach 3500 m (corresponding to the so-called "nival" vegetation belt).The Megabunus genus belongs to this tiny assemblage of high altitude species, with altitudinal records of 3200 m for Megabunus armatus (Kulczynski, 1887) and 3000 m for Megabunus rhinoceros (Canestrini, 1872) in the Alps (Beron, 2002).On the northern side of the Pyrenees (Figure 1), Megabunus diadema has a wide altitudinal range (220 -1920 m) with a likely preference for the "montane" vegetation belt (dominated by beech and fir woodland in particular; see D 'Amico & Besson, 1995), given that the mean elevation is 1143 m (median value: 1188 m).With a maximum altitude of 1920 m recorded in this study, Megabunus diadema does not reach the highest peaks of the Pyrenees (culminating at 3404 m, Pico de Aneto).
However, a deficit of field surveying beyond the tree line is not discarded.These values are narrower than those published by Merino-Sáinz et al. (2013) for geographically close populations on the southern slope of the Pyrenees (mean elevation: 1298 m; range: 800 to 2150 m) and in the Cantabrian Mountains (mean elevation: 561 m; range: 2 to 1450 m).The climatic contrast between the southern side and northern side of the Pyrenees explains this difference, through the well-known increase in vegetation belts of vegetation on the southern side (Gruber, 1978;Rameau et al., 1993) compared to the northern side.

sex-ratio
The sex ratio of a population, or a set of several local populations (whether connected or not), is a key structural variable of their natural history, as it underpins the behaviour, the social structure and the dynamics of the reproductive system (Charnov, 1982;Edwards, 1998;Frank, 1998;Hardy, 2002;West, 2009).As a result, understanding what generates sex ratio variations, is an important element to model the future growth of the population, sustainability and its vulnerability to extinction, especially when the sex ratio is asymmetrical (i.e.strongly biased in favour of one sex).
Many arachnids [including spiders -see for instance Vollrath & Parker (1992) and Aviles et al. (2000) and harvestmen -see for example Tsurusaki (1986)] show a remarkable variety of reproductive phenotypes, such as parthenogenesis, and a very strong asymmetry of the sex ratio.Megabunus diadema is an outstanding example, although available knowledge so far published mix parthenogenesis and asymmetrical sex ratio, concluding (without demonstration yet) that this limited number of males necessarily induce parthenogenesis, may this parthenogenesis be facultative.As already pointed, Philipson (1959) had described the species as parthenogenetic in England from a tertiary sex ratio of 0.25% (1 single male observed among 407 individuals), as did Stol (2005) who observed 3 males among 264 individuals (raising thus a tertiary sex ratio of 1.13 %).In the Iberian Peninsula, the recent study by Merino-Sáinz et al. (2013) reached the same finding, a highly skewed sex ratio in favour of females with a tertiary sex ratio of 13.5% (8 males observed among 59 individuals).Our study contrasts sharply with those three published studies, to the extent that we observe a skewed sex ratio in favour of males, with a tertiary sex ratio of 65.6% (486 males observed among 741 individuals).The fact that we have a higher number of individuals in the sample [741 for this study, versus 407, 264 and 59 for previously published studies: Philipson (1959), Stol (2005) and Merino-Sáinz et al. (2013) respectively] cannot alone explain the contrast.If existing, a possible bias is to be searched in the spatial and temporal variability that we characterize here for the first time, for which we describe the specific patterns.The comparison of total tertiary sex ratio (i.e. for the entire sample) as we have done for 4 populations (England, Norway, Iberian Peninsula vs northern Pyrenees) may conceal some possibly unobserved intra-populational variability and does not include the effects of altitude, time, and more importantly the interaction between altitude and time.Despite Merino-Sáinz et al. (2013) noticed that sex-ratio may locally reach a value of 30% (in the region of Huesca where 3 males and 27 females were found), our work is the only one published to date that describes the subtlety of these altitudinal and temporal variations, preventing yet further comparison.
Since the study populations in our study and that of Merino-Sáinz et al. (2013) are geographically very close, and given that the period of study is contemporary in both cases (hence excluding the possibility for any historical change in sex ratio), we advocate further studies in the whole Pyrenees mountain to understand the remarkable contrast in sex-ratios, especially in the light of recent findings.Among Arachnids, as has recently been shown for the spider Oedothorax retusus (Westring, 1851), one of the potential causes of variation in sex ratio is the level and patterns of infection by bacteria endosymbionts of maternal origin that can alter the sex ratio of offspring of a mother in order to increase their own physical condition (Vanthournout et al., 2014) The existence of altitudinal and temporal variation in the sex ratio of Megabunus diadema had never been described.This even appears to be the first documented case in harvestmen.Moreover, we show that these variations follow clear patterns: the proportion of male decreases with altitude and increases during adult development period (spring and summer).The existence of a relationship between sex ratio and altitude has already been established in other arachnids, particularly the spider Myro kerguelenensis (O.Pickard-Cambridge, 1876) in which the proportion of females increases with altitude (Lee et al., 2011).In such a case, the explanations are related to resource limitation and an increase in the prevalence of sexual cannibalism with altitude.
However, the interaction between the effect of altitude and time remains to be described because it may obscure the patterns of variation.We cannot exclude the existence of other spatially-related effects (for example a possible East/West gradient).Undoubtedly, these relationships may also be affected by differences in habitat use by females and males, as has been shown for other arachnids.These aspects deserve further investigations, especially because environmental changes quickly alter habitats and drive rapid adaptive responses by species (Bowden et al., 2015).
Revealing a preponderance of males never suspected before in this species and describing extreme variability of tertiary sex ratio values within the population (or populations) on the northern slope of the Pyrenees, we feed the debate on the mode of reproduction of the species (is Megabunus diadema really an asexual, parthenogenetic species?), and put forward the interest of this species as a model to study the underlying mechanisms in populations with a strong distortion of sex ratio.

acknowledgments
We thank those who have shared their observations of Megabunus diadema: C. Bergès, F. Blanc, S. Cally, S. Déjean, N. Gouix, L. Valladarès, y.Evenou.We are also grateful to the editor and both anonymous reviewers for their comments, which helped us to improve the manuscript.

Figure 3 :
Figure 3: Monthly variation of sex ratio in Megabunus diadema on the northern slope of the Pyrenees (n = 53 observations), from May to August.Figura 3: Variación mensual del índice de masculinidad en Megabunus diadema en la vertiente norte de los Pirineos (observaciones n = 53), de mayo a agosto.