The water mites of the genus Hydrodroma (Acari, Hydrachnidia, Hydrodromidae) in Europe and Africa

Neotypes are designated for Hydrodroma despiciens (Müller, 1776) and H. pilosa Besseling, 1940, lectotypes for H. torrenticola (Walter, 1908) and H. capensis (K. Viets, 1914). Diagnostic features and morphological variability of these two species, as well as of H. torrenticola and H. perreptans , are discussed. New morphological details are presented for H. trigonometrica Walter, 1928, H. ocellata Walter & Bader, 1952 (here considered a species incerta ), H. liberiensis Cook, 1966 and H. reinhardi Pešić, 2002. After revision of museum collections, numerous older identifications are corrected. Concerning material from Europe, a high degree of confusion is found between H. despiciens , H. torrenticola and H. pilosa . For H. despiciens , all records from Africa, as well as species identifications from Wisconsin (U.S.A) and Argentina are rejected. This species, formerly considered cosmopolitan, is actually ascertained only from the European continent. Hydrodroma pilosa is recorded for the first time from Africa (Algeria). Most published records of H. capensis refer to other, in several cases probably undescribed species; recognized material derives exclusively from the terra typica in the African Cape Province. An identification key is given for the species known from Europe and Africa.


Introduction
During early times of acarology, many species of the genus Hydrodroma (for a long time under the name of Diplodontus) or subspecies of the type species Hydrodroma despiciens (Müller, 1776) were described (see Wiles 1985). Later on, the opinion established that most of these were junior synonyms of H. despiciens. Without taking much care of character patterns and their expression in different populations, 20 th century authors tended to attribute Hydrodroma from stagnant waters to H. despiciens, from running waters to H. torrenticola (Walter, 1908). An exception was made by Besseling (1940) who was the first to detect in European stagnant waters the existence (often coexistence) of two morphological types characterized by clearly different swimming setae patterns. However, for a long time no attention was paid to his description Ecologica Montenegrina 13: 1-24 (2017) This journal is available online at: www.biotaxa.org/em of the subspecies H. d. pilosa Besseling, 1940. Wiles (1985 was the first to undertake a thorough revision of the European species of the genus, based on populations from the British islands. In view of the rather similar degree of differentiation between the three taxa in question, Gerecke (1991) proposed to rank H. pilosa at species level. Interesting information on the biology of all three species was provided by Wiles (1982Wiles ( , 1985Wiles ( , 1987, and (for H. pilosa, under the old name H. despiciens) by Meyer (1985) and Smukalla & Meyer (1987). Finally, two further Hydrodroma species were detected in Europe, namely H. reinhardi Pešić, 2002 from Balkan and the Tyrrhenian islands (recently detected also in Russia, Yaroslavl province and Caucasus : Tuzovskij 2015) and H. cf. rheophila Cook, 1967 from the Greek island Lesbos (Pešić et al. 2010). From Africa, in addition to numerous records of H. despiciens from all parts of the continent, the following species were reported: Hydrodroma perreptans (K. Viets, 1913) (Cameroon, Liberia, East Africa); H. capensis (K. Viets, 1914) (South Africa, reported northwards as far as the Sahara and also from Madagascar); H. trigonometrica Walter, 1928 (Cameroon, Algeria), H. ocellata (Walter & Bader 1952) (East Africa), H. liberiensis Cook, 1966 (Liberia) and H. zokhzovi Tuzovskij, 2014 from Ethiopia. In the course of a taxonomic study on Hydrodromidae from Madagascar, I came across with several basic questions concerning species definitions and determinations of former authors. During the attempt to redescribe Hydrodroma capensis, it became clear that numerous additional taxonomic uncertainties required a revision. The results of these studies are presented in this paper. A brief presentation of the family and genus is followed by a treatment of the European and African species, in the sequence of their description year.

Material and Methods
For morphological verifications and measurements, selected specimens were slide-mounted, partly in glycerine jelly, partly in Koenike fluid. However, during the work process it became clear that important character states such as integument structure and leg setation in Hydrodroma can easily be verified in undissected specimens. Even, possibly important features such as shape and distance of lateral eye lenses is best observed in fresh material while these characters are frequently lost in mounted specimens.

Family Hydrodromidae
The family Hydrodromidae is an isolated clade in the Hydryphantoidea. In addition to the doubtful genus Oxopsis Nordenskiöld, 1905, known from a single specimen collected in Sudan (see Cook 1974), representatives, recorded from all continents except for Antarctica, are very similar in general idiosoma shape and attributed to one single genus, Hydrodroma. Concerning many features of idiosoma and gnathosoma, hydrodromids are characterized by a high degree of homogeneity. In the following, important character complexes are briefly surveyed -for a diagnosis of the family, see Di : 1. Idiosoma integument completely soft. Except for appendages, gnathosoma, sclerotized coxal and genital plates and the unpaired excretory pore, no external sclerotization (e.g., muscle insertions, dorsal or ventral plates, secondary sclerite borders) is developed (Figs 1 A-B). The lateral eye lenses lie free under the integument and are not enclosed in capsules ( Fig. 1 A) -in contrast to most other water mites families.
2. The uppermost layer of the integument is characterized by dense papillosity. In a deeper stratum, a fine net of meshes is formed (Schmidt 1935): in top view, six finely sclerotized bars extend from the base of each papilla to form a network of equilateral triangles. Each meeting point of these lines is positioned below the tip of one of six surrounding papillae ( Fig. 1 D). Formation of the integument provides important information for species distinction: Papillae may differ in shape species-specifically, and they can be homoiomorphic (of one type) or heteromorphic (forming a pattern of two different types). In species with heteromorphic papillae, every large papilla is surrounded by six small papilla-like elevations, resulting in a honeycomb-like structure in top view (Fig. 2 B).
3. The genital field is rather uniform and without distinct sexual differences. Paired, movable halfmoonshaped plates bear a rather high number of undifferentiated, small and round acetabula and medially several rows of fine setae ( Fig. 1 B). In general, females tend to have larger genital plates and higher numbers of acetabula, but with a broad overlap in variability range (Wiles 1985). Fine sexual differences in setation were found by Meyer (1983) in H. pilosa (sub nom. H. despiciens: male with > 45 pairs, seta L generally > 100 µm; females with < 45 pairs, setae generally < 100 µm) and by Wiles (1985) in H. despiciens (male: all setae hollow; females: posterior as in males, anterior setae solid, flat, best visible under phase contrast).
4. Legs are rather uniform and, except for differences in segment size, without sexual characteristics. All leg claws are rather simple, sickle shaped, with a fine dorsal clawlet, but without denticulation or a claw blade. Differences between species may be found in the absolute and relative size of claws (larger in stream dwelling species), proportions of legs segments (stouter in stream dwelling species) and, in surprisingly stable species-specific patterns, number and arrangement of long, fine swimming setae. In addition, arrangement and shape of other leg setae, in particular those at distal margins of leg segments 3-5, may provide character states useful for species recognition.

Description:
Both sexes: Colour red, rarely orange. On each side, lateral eye lenses far distanced (> 100 µm), anterior lens round (diameter 40-50), posterior lens oval (length 40-50, width 30-45). Shape of mouthparts, coxae and excretory pore, and setation of palps as typical in most species of the genus, sexual differences very weakly developed (in general, smallest males and females similar in measurements, largest females distinctly larger than largest males, but sexes similar in proportions range); chelicera with little curved claw ( Discussion: Hydrodroma despiciens was first described by Müller (1776) from Denmark under the genus name Hydrachna and has been often misunderstood since then. The designation and detailed description of a neotype is the precondition to give taxonomic stability to Hydrodroma despiciens. The original description does not provide information on a potential type locality and I decided to select the neotype from the material collected by O. Lundblad in 1919 in the terra typica, Denmark. As typical for the consideration of this species at his time, as an ubiquist ("eurythermal, cosmopolitan"), Lundblad (1920) restricted locality information in his paper on the material in question to "generally present at all investigated sites". No collecting site list is given in Lundblad (1920), and the locality records had to be copied from his handwritten labels.
From the list of studied material above it becomes obvious that K. and K.O.Viets, as well as Lundblad did not take care of Besseling's taxonomic work and confused frequently H. despiciens and H. pilosa, K. Viets misunderstood occasionally also H. torrenticola.
In their revisional work on Australian Hydrodroma, Pešić & Smit (2007a, b, 2011) demonstrated the absence of H. despiciens from this continent and described populations previously attributed to this species as H. cooki Pešić & Smit, 2007. From a comparison between measurements taken from populations collected in Argentina attributed to H. despiciens by Cook (1980), and the data presented here and elsewhere for European populations (e.g. Wiles 1985, Tuzovskij 2015 results that also the Latin American records require revision: Specimens from Argentina represent probably an undescribed species different from H. despiciens in distinctly shorter palp and leg segments and lower numbers of swimming setae. In addition to H. capensis, in the covered area three further species are characterized by IV-L-5 lacking anterior swimming setae, namely H. liberiensis, H. reinhardi and H. zhokzovi (see below).
Biology and distribution: On the base of the set of diagnostic character states listed above, H. despiciens is defined as a stenotopic stagnant water species, characterized by leg claws relatively reduced in size (as compared with the rheophilic H. torrenticola), but clearly differing from H. pilosa with its generally higher swimming setae numbers.
For a long time, H. despiciens was considered a rare example of a water mite species with a nearly cosmopolite distribution, but its actual geographical distribution is unclear. The revision of available material shows that it is widely distributed in Central and Northern Europe, probably more scattered in the Mediterranean area. It is highly probable that it is completely absent from the African continentas also Walter confused H. despiciens with H. pilosa (see there for data from Algeria), a record from Burkina Faso (Walter 1935) is highly improbable and not documented by collection material. As the attribution to H. despiciens was found erroneous for revised material from North America, as well as for the records published from Argentina, the presence of H. despiciens in the New World is at all questionable. All American populations, partly attributed to H. despiciens, partly described as 5 subspecies of that species, require taxonomic revision. The same is true for all records from Asia - Walter (1929) and later Wiles (1986) defined H. monticola Piersig, 1896, first published from Indonesia as a subspecies of H. despiciens, as a separate species, and no record of H. despiciens is ascertained from this continent.

Discussion:
The original description was based on several specimens ("einige Exemplare"), obviously of both sexes. Only the above mentioned two specimens are found in the slide collection of NHMB, but the existence of further tube material with syntypes cannot be excluded. The lectotype male 1355/56, most probably used for the species description, is designated here in order to give taxonomic stability for the species. Walter (1908) gave a very detailed discussion of the character states of this species in comparison with H. despiciens. In agreement with his interpretation, important diagnostic features are: (1) rounded, less extended integument papillae (the presence of two types of papillae is well visible in the paralectotype, the integument of the lectotype is poorly visible due to desiccation); (2) medial setae on Cx-I not inserted on projections; (3) leg segments stouter (as documented in many L/H ratio data listed above); (4) leg claws relatively stronger (as documented by the claw L/segment 5 L ratio data above). A further characteristic feature mentioned by Walter is the reduced number of swimming setae. However, if the variability of larger numbers of specimens is considered, swimming setae numbers may overlap. As first stated by Wiles (1985), the most important difference in this regard is not a reduction, but an increase: The anterior face of IV-L-5 bears 3-4 swimming setae in H. torrenticola, but it is completely lacking such setae in H. despiciens. If variability ranges in populations of both species are considered, several measurements in the original description retained important for recognizing H. torrenticolus become insignificant: The two species overlap in size of chelicera and palp segments as well as genital plates, and they agree also widely in the number of acetabula.
Biology and distribution: First described from a spring near Naples, H. torrenticola is obviously widely distributed in Mediterranean streams (e.g. Gerecke 1991). The morphological variability of populations from Central Europe (data above and Wiles 1985), the British Isles (Wiles 1985) and Russia (Tuzovskij 2015) show no remarkable deviations from the original description and other mediterranean material. The species is thus widely distributed in Europe, but does not extend to Fennoscandia.  (11) [lacking in type]; P-2, 1.39 (17) [1.25 (17)]; P-3, 1.19 (13) [1,07 (13)]; P-4, 4.46 (41) [4,45 (41)]; P-5, 3.57 (17) [4,00 (17)  Discussion: This species was described after a single specimen of uncertain sex which is now widely lost, and there is no other material authorized by K. Viets. Most of the few specimens available from museum collections derive from areas far distant from the type localityin a geographical sense the closest specimen comes from Liberia (FMNH). Hydrodroma perreptans is extremely characteristic and cannot be confused with any species of the genus due to the widely reduced swimming setation. The specimens brought here together are surprisingly similar in important characters states and, as far as can be judged from the original description, in good agreement with the lost type specimen. Strong size differences between the small specimen from Liberia and the rather large one from Zimbabwe could result from intraspecific variability (observed to a similar degree in H. pilosa by Roy Wiles, pers. comm.), or indicate cryptic diversity.
Biology and distribution: As almost suggested by K. Viets (1914), the particular morphology of this species can be interpreted as an adaptation to life in running waters. The scattered records suggest a rather wide distribution in West, East and South Africa.     (2) "gnathosomal rostrum shorter and stouter, oral disk relatively larger": It is doubtful if this character is generally suitable for species discrimination in the genus. The investigated specimens differ from typical H. pilosa in fact as described by K. Viets, but, in contrast, in the investigated H. despiciens the gnathosomal rostrum is still more stouter and the mouth disk relatively larger. From this point of view H. capensis has an intermediate position between H. despiciens and H. pilosa, probably with an overlap with both species in intraspecific variability.

Diagnosis
(3) "cheliceral claw more strongly curved, proximal end of the basal segment more straight": The first observation is confirmed by the investigated material (compare Figs 2 F and 5 E)both, H. despiciens and H. pilosa have a straighter cheliceral claw; with regard to the shape of the proximal end of the basal segment, considerable variability is observed in the compared speciesthis part of the segment is possibly more turned up in juvenile adults.
(7) A rather strange detail is the small idiosoma size mentioned by K. Viets (L/W 870/770). Unfortunately he did not give account of the material on which his species description was based. As the two female syntypes available are distinctly larger in size, the existence of a further, now lost, smaller specimen on which the original description was based, is necessary. As in this study the idiosoma of all specimens of H. despiciens, as well as most H. capensis, is longer than 1000 µm, idiosoma size is surely not a character suitable for species discrimination in these taxonomic surroundings.
In measurements and proportions, the specimens from the surroundings of Cape Town (SMF 43882-83) agree well with the original description and data taken from the syntypes. In contrast, specimens from Congo and Madagascar which agree in simple idiosoma papillosity and II-L-5 with two swimming setae (see above, "rejected identifications (1)"), differ in a generally larger size (e.g. measurements of coxae, genital field and gnathosoma), more slender leg segments, and higher numbers of acetabula and IV-L-4 swimming setae. They are similar to Hydrodroma zhokhovi Tuzovskij, 2014, described  Diagnosis: Small in dimensions (e.g., L/W 160/123,208/168,genital plate L 157). Integument with closely arranged, flat and round papillae, hardly visible (Fig. 5 F). Cx-I medial setae not associated with tubercles; Cx-III+IV posterior apodemes long, parallel to body axis. Genital field with ca. 25 pairs of Ac, maximum number per transversal transect 3, and about 35 pairs of medial setae. Legs with relatively long claws (L ratio claw/segment 5, 10-15 %); II-L-5 with one posterior swimming seta, numerous swimming setae on III-L-4/5 restricted to posterior surface, on IV-L-4/5 on both surfaces.

Discussion:
The original description of this species was based on a single female collected by T. Monod in a crater lake in Cameroon on 22.08.1926. For the moment it remains unclear if the author was correct when he later attributed a male from Algeria to this species. It shares with the holotype several important character states (rather small in dimensions, low number of acetabula and genital setae, well developed swimming setation comprising a group of anterior setae on IV-L-5), but differs in by far stouter leg segments. As these proportions could result from squeezing of the only available male specimen, the taxonomic weight of this difference remains unclear. I decide to accept the locality record with a question mark, but not to consider the specimen for the species diagnosis.
In the original description, Walter compared H. trigonometrica with the two other Hydrodroma species recorded at that time from Cameroon, H. despiciens and H. perreptans. Hydrodroma despiciens (my revision did not confirm any African record of this species) differs, among others, in distinctly higher number of acetabula (> 50 pairs), the absence of anterior swimming setae from IV-L-5 and relatively smaller leg claws (L ratio claw/segment 5 < 10 %). As already observed by Walter, H. perreptans (and also H. ocellata) differs from H. trigonometrica among others in the morphology of legs with rather large claws (L ratio claw/segment 5 > 14, in I-L > 20 %) and an extremely reduced swimming setation (not more than one seta per segment). In general, Walter stated correctly that H. trigonometrica is a rather small species. From both compared species it differs in distinctly minor dimensions, e.g. of the coxal and genital plates, but also of palp and leg segments.
Hydrodroma trigonometrica differs also from H. capensis (and from the similar H. zhokovi) in minor dimensions and a distinctly lower acetabula number, in addition in the presence of only one posterior swimming seta on II-L-5 and the presence of a group of anterior swimming setae on IV-L-5. Hydrodroma liberiensis agrees widely with H. trigonometrica in rather reduced dimensions and a low number of acetabula, but is well distinguished by its reduced swimming setation (in total, only 10 setae, IV-L-4/5 anteriorly without swimming setae).
All things considered, H trigonometrica is obviously a valid species, but additional investigation is required for understanding its diagnostic features and their variability. The name-giving character state, the inferior integument made up of a net with triangular meshes, is obviously typically found in many, if not all, Hydrodroma species (see introduction and Fig. 1 D).
Discussion: Based on the original description, data presented by Wiles (1985) and Tuzovskij (2015), and a revision of material from various collections, H. pilosa differs from H. despiciens as follows: (1) Integument papillae of one type, distinctly away from each other, slender and sharply pointed; (2) II-L-5 with numerous (> 4) posterior swimming setae; IV-L-5 with swimming setae on both sides (anteriorly 5-9, posteriorly more than 10). In addition, H. pilosa differs from H. despiciens following Wiles (1985) in the serrulate setae at the distal margins of leg segments 4 and 5, which are higher in number and longer. A sexual dimorphism in numbers of medial genital setae reported by Meyer (1983) is supported also by data published by Tuzovskij (2015), but with a broader overlap.
The presence of numerous anterior swimming setae on IV-L-5 is also found in H. trigonometrica. This species (in parentheses) differs from H. pilosa in many features such as lower numbers of acetabula (25-30 pairs), relatively larger leg claws (L ratio claw/segment 5, 12-17 %), and minor dimensions of all parts of idiosoma and appendages. (e.g., genital plate L ♀ 155, ♂ 184). Pešić & Smit (2007a) gave a careful description of three species from Australia similar to H. pilosa in a high number of swimming setae. They found diagnostic differences in character states such as shape of genital plates, palp and leg segments, numbers of acetabula and genital setae, but also number and arrangement of swimming setae. An investigation on the formation of the integument papillae of the Australian species could give a hint on an eventual closer relationship between H. pilosa and the latter.
Locality records now available from bibliography and the revision of material from many collections suggest that H. pilosa is actually more frequent in Europe than H. despiciens and many ecological data published for the latter in reality refer to H. pilosa (see Di Sabatino et al. 2009). The pointed shape of papillae in Schmidt (1935, Figs 1-2) suggests that also that detailed anatomical treatment published under the name of H. despiciens, in reality deals with H. pilosa.

Biology and distribution:
Detailed data on the life cycle of this species were published by Meyer (1985) and Smukalla & Meyer (1987) (both under the name H. despiciens) and by Wiles (1987). Interestingly, larvae in German populations strongly preferred chaoborid hosts while populations of the British isles were also frequently found parasitizing chironomid midges. Parasitism on tipulid midges and Trichoptera was also observed. Meanwhile, the presence of this species is ascertained from nearly all parts of Europe. Concerning the tolerance of H. pilosa against electrolyte concentrations, the new records published here from nine sites in Southern Spain with a mean conductivity of nearly 3 mS/cm (for characteristics of these interesting sites see Moreno et al. 1995), support data from the Mediterranean (Gerecke 1991) and Eastern Central Europe (Kowalik 2002). In contrast, H. despiciens is more sensitive against water pollution and increased electrolyte contents and shows a preference for lower pH values (Wiles 1985). Not surprisingly, the revision of collection material shows that records under the name of H. despiciens published from (semi)arid areas in Algeria by Walter (1925Walter ( , 1928Walter ( , 1931  Discussion: Considering the extremely reduced leg swimming setation (in total only 4 short setae, located anterior on III-L-4/5 resp. posterior on IV-L-4/5), this species is obviously close to H. perreptans, not taken in consideration in the discussion of the original description. Also the integument papillosity (Fig. 4 F), as well as most measurements and the numbers of idiosoma setae and acetabula, are close to the few data available for that species. In general, the holotype of H. ocellata is a little larger in dimensions (e.g., L/W 278/193,270/245,genital plate L 225 µm). Stouter proportions of leg and palp segments are possibly due to squeezing, a proportionally large size of leg claws (ratio claw L/segment 5 L about 20 %) is found as in H. perreptans.
The name-giving condition of the lateral eyes (anterior lens bay far larger than posterior one) is difficult to observe in the holotype preparation (Fig. 4 B). Also in some of the investigated specimens of H. perreptans, both lenses differ remarkably in size ( Fig. 4 A), but this character state is not easy observed in slide mounted material: In general, the anterior lens has the shape of a football, the posterior one of a baseball, with the result that measurements may vary due to its inclination. A judgement about the shape of the integument papillae is impossible due to the bad state of conservation of the holotype.
With our present knowledge, a decision on the taxonomic state of H. ocellata is impossible. The closely related H. perreptans appears distributed over many parts of Africa, much more material is necessary in order to understand its intraspecific variability (e.g. concerning the shape of lateral eye lenses) and/or the existence of cryptic sister species.
Distribution: Only known from the type locality and four further sites in Liberia (Cook 1966) Hydrodroma rheophila Cook, 1967 (Figs 7

Discussion:
As observed by Cook (1967), this species is very similar to the African H. perreptans in the widely reduced swimming setation (Figs 8 F-H). The females from Lesbos resemble H. perreptans also in the rather large leg claws (L about 50 µm, L ratio claw/segment 5, 20-27 %), and in the heterogeneous integument structure, with larger papillae surrounded by six small, little prominent elevations. However, in the investigated specimens, the large papillae (Figs 7 C-D) are more flattened and rounded and less densely arranged than in H. perreptans. The key difference to H. perreptans is found in the stouter palp, in particular segment P-4 -L/H measurements of a lesbian specimen are: P-1, 0.8; P-2, 1,4; P-3, 0.9; P-4, 3.7; P-5, 3,7 (compare Fig. 7 E with Fig. 2 F).
Biology and distribution: As in the African H. perreptans, leg morphology and setation is an obvious adaptation to life in running waters. After the original description from India recorded also from Iran, Oman, Taiwan (with a question mark) and the island of Lesbos in the Aegean Sea (the outpost population discussed here).

Discussion:
The detailed original description does not leave important questions open, some additional morphological features could be verified from the material authorized by Pešić listed above. Hydrodroma reinhardi is similar to H. despiciens in the range of acetabula numbers (genital field: Fig. 8 B), but differs in a smooth medial margin of Cx-I (Fig. 8 A), lower swimming setae numbers (only on IV-L-4, 2 anterior setae, posterior setae numbers II-L-5, 1; III-L-4, 2-4; IV-L-5, 3-4; IV-L-4, 4-5; IV-L-5, 3-4) and relatively larger leg claws (L 15-16 % segment 5). Further differences concern generally minor measurements of palp and leg segments (e.g.,  and the shape of the integument: the papillosity consists of two different types as in H. despiciens, but the larger central papillae are not rounded, apically forming blunt tips (Figs 8 C-D).

Biology and distribution:
A running water species frequently found in springs. Central Mediterranean, recently reported also from Russia (area of Yaroslavl, Caucasus: Tuzovskij 2015); here recorded for the first time from continental Italy (Lazio).

Discussion:
The detailed original description does not leave important questions open, and no material had to be studied for comparison. Tuzovskij (2014) placed H. zhokovi correctly near H. capensis (important common character combination: presence of two posterior swimming setae on II-L-5, absence of anterior swimming setae from IV-L-5). New morphological data provided here for H. capensis confirm differences from H. zhokovi in (1) more elongate integument papillae, and (2) the genital field without remarkable sexual dimorphism, bearing lower numbers of acetabula (35-60) and medial setae (25-30). Instead, there is an overlap of the genital plate length measurement ranges, and both species agree also in the swimming setae numbers of IV-L.