1. Introduction
The genus Salamandra is distributed throughout Europe, northern Africa and the Near East. Many local populations have adapted to different climates in various habitats, but the status of species and subspecies of these populations is still a matter of dispute. Based on a mitochondrial DNA analysis, Steinfartz et al. [1] proposed the taxonomy and distribution of the Salamandra species in Europe and in Mediterranean countries. A phylogenetic analysis of the sequences suggests that six major monophyletic groups exist (S. salamandra, S. algira, S. infraimmaculata, S. corsica, S. atra and S. lanzai) that combine together more than 14 subspecies [2] . Based on the Salamandra species distribution, S. algira, S. infraimmaculata and S. corsica are apparently found at the southern border of their distribution and are exposed to semi-arid conditions [3] [4] .
Many studies have been carried out on genetic variation, not only among the species of Salamandra, but also among different populations in the same species, in order to understand the adaptation to different ecological conditions. Salamandra infraimmaculata is present in the southeastern and eastern parts of Anatolia, Turkey, northwestern Iran, northern Iraq, Lebanon and northern Israel [1] . Many aspects of S. infraimmaculata, some of which have been described in detail, have been studied over the past 40 years on the southern border of its distribution [3] [4] .
This situation is very interesting in semi-arid habitats where strong effects of the habitats were found on genetic variation, as was seen in S. algira [5] and S. infraimmaculata [3] [4] . It seems that S. infraimmaculata has been studied in more detail not only regarding genetic divergence [6] - [17] , but also in terms of other biological and ecological aspects [2] [3] [4] .
Although there are many aspects of S. infraimmaculata adaptation to dry conditions in semi-arid habitats [2] [3] [4] , relatively limited studies have been carried out on the effect of breeding places on adaptation to semi-arid habitats in S. infraimmaculata. Breeding places where larvae development and complete metamorphosis take place have been studied quite extensively, and are described in detail on the southern border of S. infraimmaculata distribution [2] [10] [18] - [26] . Relatively less information has been provided on metamorphosis and mature salamanders [27] [28] [29] .
Two different activity patterns of metamorphosis in S. infraimmaculata were described: local year-round activity [23] ; and migration to breeding places located at relatively long distances [27] . During the long migration to breeding places, the salamanders move to the same breeding places [23] [31] [32] or to new breeding places in the same area [27] .
While seeking new breeding places in semi-arid habitats where many human-made water bodies suitable for S. infraimmaculata larvae development and complete metamorphosis are located, mature salamanders and metamorphosed juveniles are unable to move back to terrestrial habitats, and later on, all the salamanders die [32] [33] .
S. infraimmaculata are facing foreseeable danger in semi-arid habitats where water bodies are limited and many human-made pits represent traps for salamanders. Estimations of damage to salamander populations are very important for nature conservation.
In the present study, the breeding places of S. infraimmaculata are evaluated and described, as well as the potential damage of pits that might represent a trap for mature salamanders.
2. Materials and Methods
Study Area
The distribution of breeding places of S. infraimmaculata in northern Israel based on maps and various studies carried out over the past 40 years is presented in Figure 1. The breeding places were divided into four categories: ponds, which is a relatively large water body where water is available all-year-round or for more than half a year; water pools, which is a relatively small body where water is available for several months during the winter; springs, in which water coming from the ground is available all-year-round; and water holes, many of which are human-made wells for collecting water.
Figure 1. The map of water bodies of potential breeding sites of S. infraimmaculata according to the maps published by the government of Israel during the years 1941 to 2018 in the area of distribution of this species in Israel. Blue represents winter reservoirs, red represents springs and streams, black represents water holes and yellow represents winter pools. The source of the map of water bodies of potential breeding sites was provided by the Tel-Hai Map Archive.
The study was carried out over eight consecutive years (2005-2014) in xeric habitats, including all habitats of S. infraimmaculata in Mount Carmel, the eastern, upper and northern Galilee, covering an area of approximately 1400 km2 in northern Israel, as described previously by Degani and Goldberg [32] , and Degani and Kaplan [10] . The habitats examined in this area were existing or potential breeding places of S. infraimmaculata (in the area of distribution of this S. infraimmaculata) where salamanders can enter but cannot exit. We focused on breeding places of rock pool holes, wells and swimming pools in the area of S. infraimmaculata breeds (Figure 2) [26] [32] [34] . We also examined potential breeding places, as well as information published in maps or on the Internet (Figure 1). Two breeding places were selected for studying both larvae and mature salamanders in detail:
1) Shazur Mountain, at coordinates 181.4/262.5 on the Israel topographical grid, an elevation of 800 m above sea level [33] . Two breeding places of S. infraimmaculata were discovered very close to this area (200 m distance between
Figure 2. Various breeding places for S. infraimmaculata including a picture of captured salamanders trying to exit the water hole but are unable to do so: A. Pond in which water is available for 5 - 10 months. B. Stream in which water flows all year round with constant temperatures (16˚C - 18˚C) all year round. C. Winter pond in which water is available for 3 - 6 months. D. Winter reservoir in which water is available during all of the winter and part of the summer all year round for 6 to 12 months. E. Spring in which water is available all year round. H. Water rock pool hole in which water is available during the winter and part of the summer all year round for 6 to 12 months. I. Water rock pool hole in which water is available in the winter and part of the summer for 6 to 8 months. J. Water rock pool hole in which water is available during the winter and part of the summer all year round for 6 to 12 months. K. Winter pool in which water is available for 6 to 8 months. L. Water rock pool hole in which water is available all year round for 6 to 12 months.
both breeding places): one in which salamanders were unable to return to their terrestrial habitat after laying their larvae (a depth of 195 cm and an area of 5 m2) (Figure 2); and the other in which they were able to return to their terrestrial habitat after breeding.
2) Maalot Rock Pool Hole, which was examined monthly over a period of four years (2005-2008). Both larvae and adults were studied. The larvae measurements were examined, as described previously by Goldberg et al. [34] . The adult salamanders were observed in the pit and outside the rock pool (Figure 2). Pictures of each adult dorsal side of the metamorphosed salamanders and measurements (weight and length) were taken [33] . The lengths and weights of adults and larvae were analyzed according to body mass index (BMI), BMI = weight/length, and one-way ANOVA, with the level of significance between groups set at p < 0.05. In addition, average lengths and weights were compared using a t-test. Each adult that was captured was photographed, its weight and length were measured [23] , and the mean ± SE of length and weight were calculated.
3. Results
S. infraimmaculata distributed in Israel are presented in Figure 1. Two different habitats must exist for S. infraimmaculata: aquatic, to enable them to complete the life cycle (see Figure 2); and terrestrial, to adapt to the salamanders’ distribution. Among the various types of breeding places, large numbers of springs and streams (503) were detected on the map of northern Israel (Table 1), and a relatively low number in water holes (48) and winter pools (25). All of these breeding sites are in the area of S. infraimmaculata distribution. In all of the breeding site types (springs, streams, water holes, winter pools and reservoirs), the larvae of salamanders were found in previous studies [2] [3] [4] [10] [18] [19] [20] [22] [23] [26] [27] [30] - [41] .
In the present study, we found that another breeding site was found for salamanders near the water hole breeding sites (Figure 3), thus enabling them to exit. In this situation, only a few mature salamanders moved to the water hole where salamanders can enter but are unable to exit (Table 2). Both males and females enter the water hole (Figure 3). During the winter, more males (42%) were observed than females 29% (Figure 4). After the salamanders’ oviposition in the water, they try to return to terrestrial habitats but are unable to do so (Figure 2), and they lose weight (Figure 5) and die. The decrease in body mass
Table 1. Water bodies in the habitats of Salamandra infraimmaculata distribution in Israel (Figure 1).
Table 2. Number of adult salamanders captured in water holes in Israel annually. Based on our detailed observation of two breeding sites, water holes into which salamanders enter but are unable to exit and die combined with the number of water holes, the estimated damage to the population of salamanders in Israel may be calculated.
Figure 3. The area around the water hole and the winter pool that were studied.
Figure 4. The composition of the metamorphosed S. infraimmaculata population (in and out of the two breeding places). The number of salamanders (S. infraimmaculata) in two breeding sizes was 57 (27 + 30).
Figure 5. The measurements of mature salamanders in winter (December) after moving to the water hole and in spring (April) when the salamanders are trapped in the pit. Each adult that was captured was photographed, its weight and length were measured [23] , and the mean ± SE of length and weight were calculated.
index (BMI) of salamanders from the winter (December) after they move to the water hole to the spring (April) when they are found in the water hole is shown in Figure 6.
4. Discussion
The contribution of this study is that it described in a relatively large area the potential damage to mature S. infraimmaculata in xeric habitats where there are limited breeding places and many water holes. Moreover, the distance between breeding places (water pools) in the present study and the water hole was at least 1 km, according the distance that S. infraimmaculata move from the location where they deposit their larvae to a terrestrial habitat. Compared to the relatively large number of studies carried out on the breeding places of S. infraimmaculata on the southern border of its distribution in xeric habitats [2] [3] [4] [10] [18] [19] [20] [22] [23] [26] [27] [30] - [41] , only a few studies described water holes as a breeding place that became a death trap for salamanders [32] [33] in the natural area of S. infraimmaculata.
The contribution of the present study is that it tried to map breeding sites and divide them into different types (springs and streams, water holes, winter pools and winter reservoirs) whereby water holes represent 8% of the breeding places. The estimation according to this study and a previous study is that there are 336 adult salamanders annually (Table 1 and Figure 2). Although many water holes have not been examined in detail in the entire area of distribution of S. infraimmaculata in Israel, death traps were found where salamanders were captured [32] [33] , some of which have not been published. Moreover, some of the new human-made water bodies, for example, swimming pools, are becoming death traps for salamanders [32] . We proposed a model for a death trap that the S. infraimmaculata would not be able to recognize during the colonization of new breeding sites (Figure 7). During reproduction, the salamanders move to
Figure 6. The body mass index (BMI) of salamanders in winter (December) after moving to the water hole and in spring (April) when the salamanders are trapped in the pit.
Figure 7. The quality model proposing the death trap in semi-arid habitats based on our studies (present study [32] [33] ).
the breeding sites. This migration in most of the semi-arid habitats is relatively high [4] [23] [30] [31] compared to moist habitats, where many streams for breeding are found [28] . In Mount Carmel, semi-arid areas of the southern population in Israel and the long distance of moving to and between breeding sites are described [27] . This study supports the hypothesis proposed by Degani et al. [33] that in salamanders, colonization is an important aspect of adaptation to semi-arid habitats, where many breeding sites are unpredictable. Some breeding sites dry up during the summer and even do not contain water in very dry years. Some human-made breeding sites become death traps and damage natural populations, particularly in semi-arid habitats where the populations are very rare. Considerably more studies must be carried out in all areas of S. infraimmaculata distribution under Mediterranean semi-arid conditions to describe this damage.
Acknowledgements
We would like to thank the Tel-Hai Map Archive for providing me with the cartographic assistance, and Dr. Alon Margalit and Dr. Salome Transkei for their help.