A new millipede-parasitizing horsehair worm, Gordius chiashanus sp. nov., at medium altitudes in Taiwan (Nematomorpha, Gordiida)

Abstract Gordius chiashanussp. nov., a newly described horsehair worm that parasitizes the Spirobolus millipede, is one of the three described horsehair worm species in Taiwan. It is morphologically similar to G. helveticus Schmidt-Rhaesa, 2010 because of the progressively broadening distribution of bristles concentrated on the male tail lobes, but it is distinguishable from G. helveticus because of the stout bristles on the mid-body. In addition, a vertical white stripe on the anterior ventral side and areoles on the inside wall of the cloacal opening are rarely mentioned in other Gordius species. Free-living adults emerged and mated on wet soil under the forest canopy in the winter (late November to early February) at medium altitudes (1100–1700 m). Mucus-like structure covering on the body surface, which creates a rainbow-like reflection, might endow the worm with high tolerance to dehydration. Although Gordius chiashanussp. nov. seems to be more adaptive to the terrestrial environment than other horsehair worm species, cysts putatively identified as belonging to this hairworm species found in the aquatic paratenic host, Ephemera orientalis McLachlan, 1875, suggest the life cycle of Gordius chiashanussp. nov. could involve water and land. The free-living adults emerged from the definitive hosts might reproduce in the terrestrial environment or enter an aquatic habitat by moving or being washed away by heavy rain instead of manipulating the behavior of their terrestrial definitive hosts.


Introduction
In addition to the two previously described species of horsehair worm (Chiu et al. 2011(Chiu et al. , 2017, Gordius chiashanus sp. nov. is the third described species in Taiwan, and one among 90 valid Gordius species reported worldwide (Schmidt-Rhaesa 2010, 2014. Gordius horsehair worms are characterized by a cuticular fold, known as postcloacal crescent, on the male tail (Schmidt-Rhaesa 2002). Gordius forms a monophyletic group (Gordiidae) with the genus Acutogordius, which bears the same characteristics; however, the phylogenetic relationship between these two genera is controversial (Schmidt-Rhaesa 2002). Although Gordius is the second most diverse genus, identification of species in the genus Gordius is difficult because of the lack of diagnostic characters and our limited understanding of its morphological variables (Schmidt-Rhaesa 2001, 2010. Phylogenetic comparison using DNA sequences with morphological descriptions has become increasingly crucial in detecting the cryptic species Tobias et al. 2017).
The definitive hosts of Gordius cover a wide range of arthropod taxa. Although many host records might be questionable because the genus Gordius (G. aquaticus Linnaeus, 1758) had been used to represent the entire members of horsehair worms, Gordius species might parasitize several insect orders, Chilopoda, Diplopoda, and Araneae as their definitive hosts (Schmidt-Rhaesa 2012; Bolek et al. 2015). The Gordius life cycle is highly correlated with the definitive hosts. The freshwater horsehair worm typically exhibits a life cycle that involves aquatic and terrestrial environments; its life cycle comprises a reproduction and paratenic aquatic host phase and a terrestrial definitive host phase (Hanelt et al. 2005). The aforementioned complex life cycle has been reported in multiple Gordius species (e.g., G. robustus Leidy, 1851 andG. difficilis Smith, 1994) (Thorne 1940;Bolek and Coggins 2002); however, it has not been reported in some species that parasitize aquatic definitive hosts (e.g., G. villoti Rosa, 1882 andG. albopunctatus Müller, 1926) (Valvassori et al. 1988;Schmidt-Rhaesa and Kristensen 2006) or in species that reproduce in terrestrial environments (G. terrestris Anaya et al., 2019) (Anaya et al. 2019).
Free-living adults of Gordius chiashanus sp. nov. are frequently found in foggy forests situated at altitudes of 1100-1700 m in Taiwan. Their taxonomic status was first examined in the present study by using a description of morphology and phylogenetic comparison of partial mitochondrial DNA cytochrome oxidase subunit I (mtDNA-COI) genes. The definitive host was determined using worms with high sequence similarity collected from the round-backed millipede, Spirobolus sp. nov. (Hsu and Chang, unpublished). Egg strings and larvae were obtained by allowing a field collected adult free-living female worm to oviposit egg string in the laboratory. The cysts which morphologically similar to the laboratory-reared larvae were collected from the field-collected mayfly naiad, Ephemera orientalis McLachlan, 1875. Based on our field observations on adult free-living worms, cysts and their hosts, along with our laboratory observations of non-adult stages for this gordiid species, we suggest the possible life history of Gordius chiashanus sp. nov.

Collection and preservation of horsehair worms
Horsehair worm samples were identified visually and collected from the ground. In total, 21 free-living adults (17 male and 4 female adults) were collected for morphological examination and DNA sequencing (detailed information provided in Table 1). All the living worms were killed by treatment with hot water (> 80 °C), fixed in a solution containing 75% alcohol with their hosts for a few days, and preserved in a solution of 95% alcohol. One mated female adult collected from Fenqihu, Zhuqi township, Chiayi county, Taiwan (23°30'12.70"N, 120°41'36.00"E) was placed in 800 mL of aerated tap water in the laboratory and maintained at 15 °C until it oviposited egg strings. The eggs were maintained in aerated water for 49 days until they hatched. One dead worm from a dead round-backed millipede (collected at 17-III-2019) and five immature worms from three of 50 round-backed millipedes (collected at 23-VII-2018 and 28-VII-2018) were collected to confirm the definitive host (detailed information provided in Table 1). All the hosts were preserved at -20 °C until dissection. The infected host and the harbored worms were preserved in a 95% alcohol solution for sequencing. Five cysts photographed from four mayfly naiads of E. orientalis collected from Lugu township, Nantou county, Taiwan (23°40'46.00"N, 120°47'18.50"E), where the freeliving adult has ever been found in the upstream of less than 1 km, were putatively identified as belonging to this horsehair worm species. All the samples were preserved in a solution of 75% alcohol for morphological examination.

Morphological examination
Free-living adults. Fragments (approximately 0.5 cm in length) of the anterior end, mid-body, and posterior end of the preserved samples were examined and photographed using a stereomicroscope (Leica S8 APO, Leica, Wetzlar, Germany), dehydrated using a series of ethanol and acetone solutions (95% and 100% ethanol (twice) and ethanol/ acetone mixtures of 2:1, 1:1, 1:2, and 0:1), dried to the critical point, coated by being sputtered with gold, and examined using a scanning electronic microscope (SEM) (JEOL JSM-5600, Tokyo, Japan) at magnifications ranging from 100× to 15,000×.  Eggs and larvae. Eggs and newly hatched larvae (living or treated with hot water (> 80 °C)) were examined and photographed on microslides by using a compound microscope (Olympus BH-2, PM-10AD, Olympus, Tokyo, Japan) at magnifications of 200× and 400×. The eggs examined using the SEM were first fixed using a solution of 75% alcohol, dehydrated, dried to the critical point, and coated with gold sputter. The eggs and larvae were examined at a magnification of 500×. ImageJ 1.47 was used for all morphological measurements (Abràmoff et al. 2004), and spatial calibration was conducted according to the scale included in each picture. The terminology for larval stages used in this study primarily followed that of Schmidt-Rhaesa (2014) and Szmygiel et al. (2014).
Cysts in the paratenic host. The mayflies preserved in 75% alcohol were first treated with Nesbitt's fluid for 15-20 min at 40 °C and a 0.1% KOH solution for 5 min at 40 °C to ensure that the cuticle and muscles had become transparent (Walter and Krantz 2009;Chiu et al. 2016). One of the cysts was further treated with a 5% KOH solution for 6 h at room temperature to release the folded larva inside the cyst wall. The cysts were examined and photographed on microslides by using the compound microscope at 200× magnification.

Phylogenetic analysis
Genomic DNA from a 1-cm mid-body section of each worm was extracted using an ALS Tissue Genomic DNA Extraction Kit (Pharmigene, Kaohsiung, Taiwan). The partial cytochrome c oxidase subunit I (COI) sequence was amplified using universal primers (LCO1490 and HC02198) (Folmer et al. 1994) or a newly designed primer set (GoCOiF-1: TTAGGAACTGCTTTAAG, GoCOiR-1: ATAGGGTCAAAGAA-GGAGG). PCR for both primer sets was initiated at 95 °C for 5 min, and amplification was conducted for 35 cycles of 95 °C for 1 min, 50 °C for 1 min, and 73 °C for 1 min, with a final extension at 73 °C for 5 min.
In addition to sequencing three free-living adult worms and six immature worms recovered from millipede hosts (242-457 high-quality base pairs), we obtained highquality CO1 sequences (>500 base pairs) from 18 adult free-living individuals to be used in our phylogenetic analysis and estimates of intraspecific genetic distances. Pairwise distance matrices of COI sequence data were calculated using the Kimura 2-parameter model. A phylogenic tree was reconstructed using the maximum likelihood method by using the General Time Reversible model with the addition of invariant sites and a gamma distribution of rates across sites. For phylogenic analysis, the COI sequences were first aligned using CLUSTALX 2.0.10 ( Thompson et al. 1997). A total of 422 base pairs shared by all the examined sequences, including for our 18 samples, Gordius/Acutogordius spp. (as reported by Sato et al. (2012), Hanelt et al. (2015), Chiu et al. (2017), and Tobias et al. (2017)) and Chordodes formosanus Chiu, 2011, Euchordodes nigromaculatus Poinar, 1991, and Parachordodes diblastus (Örley, 1881) (as reported by Chiu et al. (2011) and Tobias et al. (2017)), were analyzed using MEGA 7 (Kumar et al. 2016) (see detailed information in Table 2). One sequence of an un- determined nematomorph (MF983649) was also included because it exhibited high similarity to Acutogordius. The bootstrap method (with 1000 replicates) was used to estimate branch support of the phylogenic tree.

Seasonal occurrence of free-living adults
Seasonal occurrence of free-living adults was estimated by counting (and removing) freeliving adults (living or dead) on the ground in Dinghu, Alishan township, Chiayi county, Taiwan (23°29'29.10"N, 120°43'19.00"E) between October 2017 and May 2018.  Table 1 presents detailed information of the locality. Type material. Partial bodies of the holotype and allotype were deposited at the National Museum of Natural Science, Taichung, Taiwan. Paratypes were deposited at the National Museum of Natural Science, Taichung, Taiwan and Lake Biwa Museum, Shiga, Japan (Table 1).

Gordius chiashanus
Type hosts. Spirobolus sp. nov. (Hsu and Chang, unpublished) (Diplopoda: Spirobolidae) (Fig. 5E, F) Etymology. The specific name is the combination of chia, referring to the place (Chiayi county) where the first sample was found, and shan, referring to the Chinese word for "mountains." The word chia is also in memory of our friend, Chia-Chih Lin, who died in an accident in a field experiment.
Anterior end columnar and spherical; anterior tip white (white cap) with a dark -brown collar and a vertical white stripe on the ventral side (Fig. 1A). Under SEM, surface of anterior end appeared smooth (Fig. 1B) or wrinkled (Fig. 1C) on the tip of one sample; scattered short bristles (11.24 ± 6.57 (4.92-22.24) µm in length) scattered except on tip in most samples (Fig. 1B, D).
Phylogeny. The partial COI sequences of the 18 free-living adults contained 15 haplotypes with 392 invariable sites, nine singletons, and 21 parsimoniously informative sites. The genetic distance among them was 0.0024 within the range of 0.0000-0.0510. The three living adults and six worms inside the hosts were considered conspecific with the 18 free-living adults because of their small genetic distances (0.0000-0.0719). The mean interspecific genetic distances between Gordius chiashanus sp. nov. and other Gordius species or clades were in the range of 0.2320-0.4242, and that between Gordius chiashanus sp. nov. and Acutogordius taiwanensis was 0.3648 (Table 3). In addition to short genetic distances, the conspecific status of the 18 free-living Table 3. Intra-and interspecific mean COI genetic distances of Gordius/Acutogordius species or clades under K2P model. -Indicates a single haplotype whose intraspecific distance could not be calculated. Figure 7. Phylogenetic relationship of Gordius/Acutogordius spp. restructured using COI partial sequences compared with C. formosanus, E. nigromaculatus, and P. diblastus as out groups. Numbers at the nodes represent the percentage of 1000 bootstrap replicates. adults was also supported because all the samples were located in a single clade, as indicated by a high bootstrap value. No subgroup was detected because the polytomic topology exhibited low bootstrap values and short genetic distances. The Gordius species/clades in the present result were consistent with the results of Hanelt et al. (2015) and Tobias et al. (2017), despite slight differences in the relative relationships among species, which might be attributable to the differences in models used or the shorter sequence adopted in previous studies. The clade of A. taiwanensis was located within that of the Gordius species, and it did not behave as a sister group (Fig. 7).
Reproductive season. Free-living adult worms frequently aggregate and mate on wet ground (Fig. 5B, C) after rain or fog, and they are sometimes found in water or soil (Fig. 5D). They suddenly emerge in early December, and their number decreases within 1-2 months (Fig. 8). During the reproductive season, no infected host was found. The seasonality and pattern of Gordius chiashanus sp. nov. differed from the graph constructed using data from C. formosanus (Chiu et al. 2016).
Diagnosis and comments. The 21 free-living Gordius adults and six juvenile worms from round-backed millipedes were judged as belonging to the same species in accordance with the results that they all were located in the same clade in the phyloge-netic tree and had low genetic distances (Fig. 7, Table 3). These samples were regarded as a new species, Gordius chiashanus sp. nov., on the basis of their distribution patterns of bristles on the male tail and presence of a vertical white stripe on the anterior ventral side and areoles on the inside wall of the cloacal opening.
The concentration of bristles and spines on the male tail lobes has been previously described in species from the Palaearctic (Spiridonov 1984;Schmidt-Rhaesa 2010) and Nearctic realms (Anaya et al. 2019). In Gordius chiashanus sp. nov., this dense patch of bristles is a stable characteristic that was detected in all samples. The distribution pattern was similar to that of G. helveticus (Schmidt-Rhaesa 2010) because the bristles exhibited a progressively broader distribution instead of being concentrated along the row of the ventral border, such as in G. karwendeli Schmidt-Rhaesa, 2010 (Schmidt-Rhaesa 2010) and G. terrestris (Anaya et al. 2019), or in a circular patch of concentrated spines, such as in G. spiridonovi Schmidt-Rhaesa, 2010 (Spiridonov 1984).
Although the distribution pattern of the bristles is similar to that of G. helveticus, G chiashanus sp. nov. is morphologically distinct because of the presence of stout bristles on the mid-body, a vertical white stripe on the anterior ventral side, and areoles on the inside wall of the cloacal opening. The vertical white stripe on the anterior ventral side can be easily observed by the naked eye, but it has rarely been mentioned thus far. The presence of a white stripe was previously reported in the terrestrial hairworm, G. terrestris (Anaya et al. 2019), which exhibits a broad white patch; however, the patch is likely to be the intensive aggregation of white spots in Gordius chiashanus sp. nov. The presence of areoles on the inside wall of the cloacal opening has only been reported in an unknown Gordius (Schmidt-Rhaesa 2012, fig. 3.2.2). Although cloacal openings are usually covered by contamination in many Gordius species, as was the case in most of our samples, the areole on the inside wall of cloacal opening might not be a general characteristic of the genus Gordius because it is absent in at least some species (e.g., G. serratus Schmidt-Rhaesa, 2010, G. terrestris, G. spiridonovi) (Schmidt-Rhaesa 2010Schmidt-Rhaesa and Prous 2010;Anaya et al. 2019).
Morphology of Gordius chiashanus sp. nov. With approximately 90 valid species, Gordius is the second most diverse genus of the phylum Nematomorpha (Schmidt-Rhaesa 2012). However, because of the lack of reliable diagnostic characteristics and non-hereditary morphological variation associated with methods of examination, environmental damage, mucus-like structure covering the surface, and different hosts, species identification within this genus is difficult (Schmidt-Rhaesa 2001, 2010Chiu et al. 2011Chiu et al. , 2017Hanelt et al. 2015). Previously, the white spot has been only found on the male cuticle (Schmidt-Rhaesa 2010). However, we found it, but unexpectedly, in the female Gordius chiashanus sp. nov. by examination with a compound microscope. It is clearly necessary to reexamine other species since it might have been ignored especially in the female samples. The mucus-like structure is the structure covering the body surface which might also cause morphological variation. It was first reported in A. taiwanensis (Chiu et al. 2017) but not in our observations of C. formosanus (Chiu et al. 2011(Chiu et al. , 2017. In Gordius chiashanus sp. nov., it was more obvious than that of A. taiwanensis by the bright light reflection on the body surface and the hazy appearance surrounding the worms after treatment with hot water. The mucus-like structure appeared opaque under the SEM; this opacity might hamper the visibility of small structures (Fig. 1C, D), consequently, the reliability of such a diagnostic characteristic is low.
Adult and larval size. The body length of Gordius is variable and can be longer than 2 m (Schmidt-Rhaesa 2010). Relative to phylogeny, host size and intensity of infection play more crucial roles in determining worm size (Hanelt 2009;Chiu et al. 2017). Although the adult length is less likely to be a common feature shared among a species, larval size might have been overlooked. Hidden diversity due to large cysts in the paratenic host is often detected (Chiu et al. 2016). Larvae of Gordius chiashanus sp. nov. are morphologically similar to A. taiwanensis (Chiu et al. 2017) but significantly longer than A. taiwanensis larvae (preseptum + postseptum: 162.80 ± 1.78 µm vs. 112.00 ± 5.52 µm, larvae treated with hot water). In terms of comparison with other Gordius species, although the measurements varied considerably among the untreated larvae, the larval lengths of Gordius chiashanus sp. nov. (115.38 ± 12.08 µm) were similar to those of G. cf. robustus # 1 (110.0 µm in Szmygiel et al. 2014) but longer than the unfolded larva of a Gordius species (80.02 µm in Fig. 1D, Harkins et al. 2016) and shorter than those of G. cf. robustus # 2 (140.2 µm in Szmygiel et al. 2014). The fine structures of larvae are potential to be adopted in distinguishing the close species. By examining with SEM, Anaya et al. (2019) found differences in the number of spines on the proboscis, while G. terrestris has seven spines on the distal end of the left lateral and right lateral sides, whereas there are nine in G. cf. robustus #1 (Szmygiel et al. 2014). Similarly, the pattern of spines on the proboscis is also different in C. formosanus (nine on the distal end of the dorsal and ventral sides (Chiu et al. 2011)) and C. morgani, C. kenyaensis, and C. janovyi (5 on the each side) (Bolek et al. 2010(Bolek et al. , 2013Szmygiel et al. 2014). In this study, larvae of Gordius chiashanus sp. nov. are failed to be examined by SEM, but it is worth to compare the larval morphology through the horsehair worm species in future studies.
Phylogenetic relationship of Gordius and Acutogordius. Molecular comparisons have been rarely conducted in the 19 nematomorph genera (Bleidorn et al. 2002;Efeykin et al. 2016), and the present study is the first examination of the phylogenetic relationship of Acutogordius and Gordius belonging to the family Gordiidae. Because of the shared characteristic of the postcloacal crescent, Acutogordius was considered to be phylogenetically close to Gordius but distinct because of its pointed tail lobes (Schmidt-Rhaesa 2002). Two hypotheses have suggested that Acutogordius might act as a sister group or a subtaxon of Gordius (Schmidt-Rhaesa 2002). Our results indicate that the genus Acutogordius is a subtaxon of Gordius species, although including only one Acutogordius species in analysis is insufficient to support a monophyly of the genera Gordius and Acutogordius. Moreover, our results suggest that Acutogordius might be a group of Gordius that adapts to tropical habitats. The three clades of tropical horsehair worms are grouped together with the sequences for A. taiwanensis from Taiwan, one sequence from Myanmar (Myanmar nematomorph, MF983649), Gordius sp. N178 (KM382321) from Nicaragua, and Gordius sp. N178 (KM382322) from Malaysia. The adaptation to the tropical habitat of these two genera corresponds with the global distribution. Acutogordius species are mostly distributed in the lower latitude regions; by contrast, the Gordius species mainly inhabits the Palaearctic realm (Schmidt-Rhaesa 2002, 2014Schmidt-Rhaesa and Geraci 2006;Schmidt-Rhaesa and Schwarz 2016;Chiu et al. 2017). In addition, similar patterns were observed in the altitudinal distribution of these two genera in Taiwan. Acutogordius taiwanensis mainly inhabits low-altitude rivers (Chiu et al. 2017), whereas Gordius chiashanus sp. nov. is only found in mountains at 1000 m. It is worth to note that Gordius chiashanus sp. nov. is in the same clade with G. terrestris and G. cf. robustus (clade 8). Despite not highly supported by the bootstrap method, these three species show a distinct similarity in biology. The definitive host of G. cf. robustus (clade 8) is the millipede, whereas that of most of G. cf. robustus (clade 2, 3, 4, 6) are orthopterans . For Gordius chiashanus sp. nov. and G. terrestris, the egg with a distinct membrane around the larva and the free-living adapting to terrestrial environment have never mentioned in other species. This clade of Gordius might represent a unique life history of the horsehair worm.
Definitive host and route of transmission. The millipede has been known to be the host of horsehair worms, including the genera Gordius and Gordionus (Schmidt-Rhaesa et al. 2009;Schmidt-Rhaesa 2012;Hanelt et al. 2015). As a detritivore, it is less likely to ingest horsehair worm cysts from the paratenic host. In 1930, Dorier suggested water and vegetation possible route of transmission after observing the formation of horsehair worm cysts in the external environment instead of inside the paratenic host (reviewed in Schmidt-Rhaesa et al. 2009). Recent observations of free-living cysts support this hypothesis Chiu et al. 2017). However, a detritivore definitive host can also be infected by ingesting corpses of the infected paratenic hosts. The cysts, which were putatively identified as Gordius chiashanus sp. nov., found in the mayfly naiads suggest that this is a possible route of transmission. However, the prevalence was low (3.85 and 8.33% from 26 and 24 hosts collected in Shihjhuo in the end of July). It might suggest the less efficiency in transmission or the under estimation of the prevalence since the samples were collected 4 months before the worm appeared on the soil surface.
Host and host manipulation of horsehair worms. The host and biological characteristics of Gordius chiashanus sp. nov. suggest an atypical life history. In general, freshwater horsehair worms (gordiids) develop in terrestrial definitive hosts and reproduce in aquatic environments (Hanelt et al. 2005). Adult worms maturing in terrestrial hosts have long been observed and confirmed through experimentation to manipulate host behavior to facilitate host falling into water, which enables them to reproduce in water (Thomas et al. 2002;Sanchez et al. 2008;Ponton et al. 2011). However, these observations are confined to the gordiids parasitizing a few host taxa (mantids and orthopterans) (Schmidt-Rhaesa and Ehrmann 2001; Thomas et al. 2002), whereas that parasitizing other hosts, crossing several arthropod taxa (Schmidt-Rhaesa 2010; Bolek et al. 2015), is likely to exhibit the different reproductive strategy. The alternative nonmanipulative hypotheses include the "chance hypothesis" suggested by observations of adult C. ferganensis Kirjanova & Spiridonov, 1989 emerging from mantids that drowned in small puddles formed by heavy rains (Kirjanova and Spiridonov (1989), reviewed by Schmidt-Rhaesa and Ehrmann (2001)). The "aquatic life cycle hypothesis" is suggested by the Gordius spp. parasitizing aquatic caddisfly larvae as definitive hosts (Valvassori et al. 1988;Schmidt-Rhaesa and Kristensen 2006), and the "terrestrial life cycle hypothesis" suggested by G. terrestris laying eggs in wet soil (Anaya et al. 2019).
In this study, the female adult oviposited in the water. The cysts found in the aquatic paratenic hosts and the eggs developing in water also suggest the life cycle of Gordius chiashanus sp. nov. could occur in water and on land. However, the current evidence did not exclude the oviposition in the terrestrial environment because no terrestrial paratenic host was examined for cysts. In addition, the double membraned egg (Anaya et al. 2019) and the mating on the ground both suggest Gordius chiashanus sp. nov. might be able to reproduce in the terrestrial environment. Regardless of the scenarios, the adult worm might not be carried to water by manipulating behavior of its millipede host. Alternatively, they may emerge in the terrestrial environment, and move into the water or reproduce in the soil. Free-living adults of Gordius chiashanus sp. nov. are frequently found moving and mating on the surface of wet soil during periods of fog and rain. The mucus-like structure, which causes a rainbow-like reflec-tion, might endow the worm with a high tolerance to dehydration. In the winter (late November to early February), the number of free-living adults sampled from the surface of the soil, suddenly increased and then steadily diminished. The adult C. formosanus has a pattern that differs from the bell curve in terms of its presence inside a manipulated host (Chiu et al. 2016, fig. 8) and free-living adults of G. difficilis in the water (Bolek and Coggins 2002). This difference suggests that the seasonal occurrence of Gordius chiashanus sp. nov. does not represent the time when the worm matures but the time of reproduction after the free-living adult has waited for suitable soil conditions. That worms emerging from the hosts in the soil might explain why infected millipedes are rarely found on the ground.