This is the first time mermithids have been found in the genus Coelotes (Agelenidae) and Piratula (Lycosidae), and in the lycosid species Alopecosa pulverulenta and Pardosa paludicola. Concerning Agelenidae, which are typically funnel web weavers, the mermithid was only found in Agelenopsis oregonensis Chamberlin & Ivie, 1935 [10]. Our finding comes from the genus Coelotes, most probably Coelotes terrestris (Wider, 1834), which is very common in the area and was recorded at the same site from adult specimens. Spiders from this genus are ground dwellers and they are regularly sampled by pitfall traps. The representatives of Lycosidae, which are ground hunters, are most frequently recorded as hosts of mermithids. The nematodes have been recorded in 23 species belonging to 10 genera of Lycosidae, which constitute 37% of all known spider hosts. This applies in particular to the genus Pardosa (13 species, 20% known spider hosts) [6,30,31]. Alopecosa as mermithid hosts have only so far been known from A. inquiline (Clerck, 1757) and A. trabalis (Clerck, 1757) [10]. We add A. pulverulenta to the list, a very common spider species. Concerning the genus Trochosa, the nematode parasitoids were only found in Trochosa robusta (Simon, 1876) [32] and Trochosa cf. terricola [33]. Our record comes most probably from Trochosa spinipalpis or T. terricola, which prefers areas with high humidity. Females from the genus Trochosa in Central Europe are hard to identify to species level based on the structure of their genitalia, however it is not impossible [23]. The specimens we observed strongly resembled T. spinipalpis, nevertheless some doubts remained. Nevertheless, T. terricola was also present in the same peat bog, and the proportion of adult male numbers was 5.4 : 1 in favour of T. spinipalpis.
Mermithid parasitism of spiders is rather widespread, although rarely recorded [6,10]. Only in a few studies was estimation of the prevalence of nematodes possible. In our study the rate of parasitism in the Pardosa paludicola population was up to 10%. Similar results, indicating a high infection rate, were recorded in a population of a ground dwelling mygalomorph spider - Atypoides riversi O. P.-Cambridge 1883 in California [10] and in a lycosid Pardosa milvina in Illinois [34]. In the latter case, 8% of individuals were infected by mermithids. A lower prevalence of parasitoids was recorded in a population of Pardosa glacialis (Thorell 1872) in Canada, in which 0-5% individuals were parasitized, and the result depended on the changing humidity of the areas along the stream [10]. A similar level of infection rate (from 0 to more than 4%) was noted in long-jawed orb-weaver spiders Tetragnatha (T. brevignatha Gillespie, 1992; T. quasimodo Gillespie, 1992; T. anuenue Gillespie, 2002) in Hawai [17]. Mermithids are usually noted as they emerge from a spider and little is known about their development in spiders. A relatively complete life cycle is known only for Aranimermis giganteus, parasitoid of mygalomorph spiders [12]. In the case of this mermithid, however, it is common that more than a single individual develops in a host (up to seven specimens from a single host) [12], whereas in most of the Araneomorpha spiders only a single mermithid appears in each host (e.g. [6,15]). The only exceptions were observed in two specimens of Pardosa pseudoannulata (Bösenberg & Strand, 1906), from which two and ten parasitoids emerged [35].
Many authors pointed out that the indirect life cycle (involving a paratenic host) might be more common for mermithid recorded from spiders, especially taking into account that mermithids were recorded both from active hunters and web spiders [10,12,15,35]. On the other hand, the parasitized spiders are mainly recorded in wet habitats [31,35,36], which may be suitable for mermithid pre- and postparasitic stages. A large number of known spider hosts (about 41%) are ground dwellers [6], thus it seems possible that both life cycle types (direct and indirect) could be involved in the transmission of spider parasitoids belonging to Mermithidae [10].
In Pardosa paludicola, mermithids emerged near the epigyne, in the same way as in Tenuiphantes sp. [14], but in Piratula hygrophila – similarly to Trochosa sp. – the parasitoid came out near the spinnerets [33]. Before the emergence of a parasitoid most of P. paludicola showed no external signs of parasitism, with the exception of swollen opisthosoma, which was of a similar size to those of unparasitized females before making an egg sac. An enlarged opisthosoma was also observed in parasitized adult lycosid males of Prolycosides amblygyna (Mello-Leitão, 1942) and juveniles salticids (Thiodina sp., Frigga sp.) [31]. In infected females of Pardosa pseudoannulata, the opisthosoma was swollen and sometimes lopsided [35]. On the other hand, no external morphological change was observed in parasitized adult females of Heriaeus spinipalpus Loerbroks, 1983 [36] or Caerostris sumatrana Strand, 1915 [6]. The occurrence of swollen opisthosoma can be a result of the length of the mermithid and the body size of spiders [6,11], but also the placement of parasitoids in the host. In one of the parasitized females of Pardosa paludicola, a torsion of the pedicel was noticed after the spider’s death. After the spider dissection it was revealed that the parasitoids had penetrated partly into the prosoma by the pedicel. This observation corresponds with the observation by Ranade & Prakash [37] that a mermithid was found in the cephalothorax of Heteropoda venatoria (Linnaeus, 1767). Leech [13] showed that a parasitized Pardosa glacialis lacked some of the main prosomatic muscles, which may also suggest that mermithids in this species also resided in the prosoma.
In the opisthosoma, mermithids feed on the spider's tissues and haemolymph [38]. Leech [13] observed a lack of the entire digestive system, and body fat as well as various degrees of castration in infected males and females of Pardosa glacialis. The latter was also noted in an Alopecosa inquiline (Clerck, 1757) [39]. Castration can lead to anomalies in the copulatory organs. The maldevelopment of the epigyne, the external, sclerotized plate in spider females that is used for copulation and oviposition, may be caused by several factors [40] including teratologies [41,42], hybridization [43], development of gynandromorphic specimens [42,44], and damage of genitalia during mating [45], as well as parasitoids. Epigyne deformation caused by a mermithid was documented in Pardosa furcifera (Thorell, 1875) [11], Pardosa glacialis [13] and Philodromus collinus Koch, 1835 [18]. In the latter case this even led to the description of a new species, based on the undeveloped epigyne or pre-epigyne of infected specimens [18,46]. The undeveloped spermathaecae in parasitized Piratula hygrophila that we described could lead to a similar mistake. In general, the modifications of spider genitalia as a result of mermithid infection are extremely variable. We observed a whole continuum of changes, from a largely modified epigyne, which could easily be assigned to a completely different species, up to no visible modifications of external genitalia. The same rule applies to the data from the literature. Parasitoids do influence the maldevelopment of genitalia in some cases, however the extent of change might be a result of different, still unknown factors, e.g. time of infection, placement of a parasitoid, parasitoid species, and host organs that are first eaten out during the infection.
Different parasites and parasitoids may modify the behaviour of the spider [47]. In the case of mermithids there is a report that infected spiders move towards sources of water, where the parasitoids emerge from the bodies of the hosts to complete the life cycle [10,12]. Altered behaviour was shown by Cantuaria females infected by Aranimermis giganteus, which were found in a pan trap, whereas it is unusual for the female of this spider to leave or move far from its burrows [12]. In the case of our observations from the Izera Mountains, the spiders with nematodes were mostly found in pitfall traps that had been flooded (almost totally filled up with water). The peat bog in which parasitized Trochosa were observed was inhabited by two species from this wolf spider genus. Although all the mires were dominated by Trochosa spinipalpis, we cannot exclude that specimens of Trochosa terricola, which are common, and affiliated to drier habitats than T. spinipalpis, migrated more eagerly from the adjacent habitats due to changed behaviour. Only one parasitized P. paludicola kept in a laboratory showed behavior change. This female turned upside down two days before death, which could have been the result of torsion of the pedicel caused probably by the presence of a mermithid in this part of the body. In Pardosa glacialis, when the parasitoid was about to emerge, the spider crawled into a dark hole or corner [13].
As a parasitoid, the mermithid leads eventually to the death of the host. In eight out of nine cases of parasitized Pardosa paludicola, hosts were found dead during or after mermithid emergence. Because spiders were checked twice a day, it is possible that infected spiders died similarly as in most other cases, for example Pardosa glacialis died 30-60 minutes before a mermithid emerged [13], Pardosa pseudoannulata was usually dead at the time of nematode emergence [35], Caerostris sumatrana Strand, 1915 died immediately [6] and Heteropoda venatoria (Linnaeus, 1767) within an hour after parasitoid emergence [37]. It was a huge surprise that one female of P. paludicola lived for seven weeks after the emergence of the parasitoids. Two salticid spiders (Thiodina sp. and Frigga sp.) [31] and individuals of Pardosa milvina [34], also remained alive, but only for a few days or 24 hours after parasitoid emergence, respectively.
Up to now, only five sequences of 18S rRNA from mermithids emerged from spiders have been known – three from Tetragnatha sp. (Tetragnathidae) from Hawaii [17] and two from Philodromus collinus (Philodromidae) from Germany [18]. Unfortunately none of them create a phylogenetic branch together with the haplotypes of studied specimens. All known haplotypes of mermithids obtained from spiders form clusters with haplotypes belonging to specimens found in insects, or even crustaceans (Fig. 5). Our results show that different mermithid taxa (species and genera) infest spiders, and that the mermethid parasitoid of spiders arose independently in different phylogenetic branches of Mermithidae.