Does the African native host explain the African origin of the parasite? The Maltese Geckobia estherae n. sp. parasitic on Tarentola mauritanica (Acari: Raphignathoidea: Pterygosomatidae)

A BSTRACT — As in other geographic situations, Maltese populations of the common Mediterranean wall gecko, Tarentola mauritanica , were found to be simultaneously infested by two species of geckobians, Geckobia loricata and Geckobia estherae n.sp. . The latter species exhibits modiﬁed scale-like setae. The geckobians, protected ventrally by such scale-like setae, are widely distributed: in and around the Mediterranean Basin, these species were described notably on Phyllodactylidae ( Tarentola ). Most of them belong to the group I deﬁned by Jack (1964) considering the leg chaetotaxy. The distribution of geckobians allied to the new species is discussed and correlated to the most recent knowledge on the genus Tarentola . The host being primarily African, the possible African origin and the dispersion by wall geckoes of these geckobians is discussed. As far as we know, such scaly species are not present in South or North America. The acquisition of this adaptation should therefore date from the early Cenozoic. The authors conclude that new investigations are necessary to examine this hypothesis on the Mediterranean gekkotans.


INTRODUCTION
"Delineating species boundaries is crucial because it is the first step toward discussing broader questions on biogeography, ecology, conservation, or evolution. The gap in communication between different disciplines currently related to species recognition is an important but often overlooked problem" (Carranza et al. 2002).
With few exceptions, the Pterygosomatidae (often called as "scale mites") are specialist parasites of reptiles. These blood suckers spend more than 90% of their lifetime fixed deeply by their chelicerae on the host's teguments. If the most primitive genera are not greatly modified by the parasitic way of life (Pimeliaphilus, Hirstiella and Geckobiella), the others have integrated the constraints of life on a type of host and are highly specialized: the diverse Geckhttp://www1.montpellier.inra.fr/CBGP/acarologia/ ISSN 0044-586-X (print). ISSN 2107-7207 (electronic) Bertand M. et al. oes are targeted by the genus Geckobia, the Agamidae by Pterygosoma spp.
In this article, the new Maltese species Geckobia estherae n. sp. found on the common wall gecko is described. It was the opportunity to complement (and to interrogate on) the larger question of the links between the genus Tarentola L., 1758 and the geckobians in the Mediterranean Region. Recently, the phylogenetic classification of Gekkotans integrated new analysis and discriminates now the Phyllodactylidae (i.e. genera Tarentola, Phyllodactylus) from the Gekkonidae (Gamble et al. 2008a(Gamble et al. , 2010. These two families diverged during the Mesozoic (Vidal and Hedges 2009). The genus Tarentola (around 20 species) belongs to the Phyllodactylids. All species are robust, without divided subdigital lamellae, and well-developed claws on the third and fourth digits and, with a unique exception, have a conservative morphology (Carranza et al. 2002). The genus Tarentola sensu stricto is distributed in North Africa, southern Europe, and the eastern Canary Islands. The Linnaean species "Tarentola mauritanica" can be regarded now as a complex of species and subspecies with an African origin (Harris et al. 2004, Jirku et al. 2010; the parasitic geckobians have been described several decades ago, and by tradition, the two species collected on it are identified as Geckobia latastei Mégnin, 1878 and G. loricata Berlese, 1892(Haitlinger 2004, Willmann 1955. However, new endemic geckobians were described on neighbor and insular host species, notably in Canary Islands (Zapatero-Ramos et al. 1989).
The collect of the new species in Malta is an original data. It permitted to verify: (i) that two geckobian species are hosted, by Tarentola mauritanica in different part of its distribution, (ii) that the two species, as in other situations, differ greatly each other, (notably by presence or absence of special shaped ventral hairs), and (iii) to interrogate on the correspondence of endemic hosts and the consequences for parasites around the Mediterranean Basin.
Two new questions are expecting answers, consecutively to the new data on host's origin and diversity: i) Why is it possible that, repeatedly, two so similar parasites share the same host? May be because the parameters of infestation by two species respectively with normal and scale like ventral setae differ? This problem was already investigated by Girot (1969).
ii) Can we reckon that the presence of highly modified ventral setae reveals a possible common origin of the Geckobians exhibiting this character, at least among the species parasitizing the host Tarentola spp.?

MATERIAL AND METHODS
Mites were collected on one male and two females (altogether 3 specimens) of the hosts, on the back of their heads, with spraying a small amount of surgical spirit onto the geckos body and then removing mites with a pincer. Some mites were cleared in lactic acid, and dissection of mouthparts and legs were done. Microscopic observation was helped by a microscope Wild Leitz 20 EB, measurements (in µm) were made on pictures taken with calibrated Motic camera and drawings were made using a camera lucida. Nomenclature follows Jack (1964), and Bertrand et al. (1999).
The Maltese herpetofauna: The only two Maltese geckoes (called "Wizgha" in Maltese language) are Hemidactylus turcicus and the Maltese wall gecko T. mauritanica. The present herpetofauna list of the Maltese islands consists of only sixteen species, of which five are marine turtles. Besides the mentioned geckoes, the terrestrial reptile species are four species of Colubridae, one Chamaeleontidae, one Scincidae and one endemic Lacertidae, with several subspecies. Two Amphibian species occur, a native frog from the Discoglossidae and one introduced species from the family of Ranidae (Sciberras, 2007). The only possible hosts are the Phyllodactylidae Tarentola and the Geckonidae Hemidactylus.
The host belongs to the genus Tarentola, the Mediterranean geckos, widely distributed. It is an African native genus: "Tarentola is generally very uniform and may have been so for over 10 million years; this is not due to any overwhelming phylogenetic constraint" (Carranza et al. 2002). Around the Mediterranean Sea, Tarentola is represented by several species and some subspecies have also been described: to date, the nominative subgenus comprises several endemic species: Tarentola  Female -Twelve specimens. Body roughly triangular in shape, red colored, wider than long, rather small in size, length and width of animals preserved in ethanol: 260 -290 long, 480 -490 wide. Soft cuticle, covered by long and scarcely dense setae on the dorsum ranging from 25 µm to 75 -80 µm in length.
Dorsal view (Fig. 1a) -Dorsum: cuticle with striation, covered by pilose setae (ca. 100 dorsal without scutal and lateral setae) longer on the central and peripheral parts than on the anterior. The scutum (Figs. 1a, 2a) is symmetric, clearly wider than long, generally with incised backward arbelos shape, centrally concave backwards. The scutum does not reach the slightly concave anterior edge of the body. Posterior to the frontal edge of the scutum, a double pair of setae is present. Ocular lenses are situated on a small latero-anterior extension of the scutum. Three paired lateral scutal setae, placed in rows all along the lateral limit of the scutum, four pairs of mediolateral setae and four pairs of setae along the concavity of the posterior limit of the scutum are present.
Dorsal setae, (more than 60 excluding peripheral and scutal setae) fairly elongated, are arranged in files symmetrically on each side of the body with a four setae file (each 60 -75 µm long) in the median part of the dorsum. Medially the anogenital fields are visible with three short anal setae (30 -45 µm) and three longer perigenital setae (75 µm in length) (Fig. 1d).
Gnathosoma -Infracapitulum: In ventral view (Fig. 1c), the infracapitulum is subquadrangular in shape with one pair of long setae (= gnathobasal setae) slightly longer than the palp that can reach 100  µm length. In dorsal view, the pedipalps (ca. 100 µm long), the chelicerae (>120 µm) and the lips tube are visible, flanked by the peritremes, directed upward. Chelicerae are 100 -120 µm long. The proximal end (30 -40 µm) is subquadrangular in shape; it continues in a long slender branch ended by the mobile digit abruptly ended by 1 -4 little teeth protected by an incomplete hood (Fig. 2b). Palpi: femora are as long as wide, the femoral dorsal setae are ciliate, shorter than the dorsal seta of genua that is simple. Palpal chaetotaxy is 1-1-2-4 + ω. Dorsal tibial seta is rather long compared to other species, but distinctly shorter than femoral and genual ones. The two longest setae of tarsus are 25 -30 µm long. The terminal claw is rather long, embedded and protected by a developed cuticle fold.
Legs -The shorter legs are the leg I and II (140 -160 µm long) the longer are III and IV (200 -250 µm long). Separations between coxae I and II and coxae III and IV are not visible. Only the setae on coxae I are simple and long (40 -60 µm). On some specimens, the distal seta on coxa I was not observed. Other setae are stout and short (ca 25 µm long). Coxal chaetotaxy= 4 or 3(1 or 2+2), 5(2+3). Anterior and posterior epimera are well separated and a file of two setae similar to the stout setae of coxae II and IV are situated on this inter-epimeric space.
Tibial, genual, and femoral setae are simple except on the fourth pair of legs; this ventral femoral seta is plumose. All trochanteral setae are plumose and strong.
Male -no male was collected.
Etymology -Species named in honor of Mrs. Esther Sciberras who assisted in the collection of the specimens and data during the field works and supported the herpetological works of Mr. Sciberras for years.
3. The femoral palpal seta is long and ciliate, differs from the stronger and ciliate ones in the species described from Canary Islands; robust, slender and ciliate, it is less long in G. tarentulae and G. loricata. The ocular sclerites are fused anterolaterally to the scutum in prolongation with the anterior angles of the scutum (individualized from the scutum in the Canary species and G. loricata). The ventral scalelike setae are "crenellate" as in G. turkestana (Hirst 1926), but neither heart-shaped (G. loricata, G. turkestana) nor fan-shaped (G. tarentulae), nor elongated in length (G. hindustanica Hirst, 1926). They are replaced with shortened setae (brush-like) from the center of the venter towards the lateral sides. As previously observed on several species (G. turkestana, G. loricata, G. squamaeum Bertrand et al. 1999and G. parvulum Bertrand et al. 1999: Hirst 1926Girot 1969;Bertrand et al. 1999), the shape of the peripheral setae changes gradually to become more and more elongated, first to "leaf-like setae" before making place to normal, "seta-like setae". A proximal tomentous and stout seta and a second, shorter one distally are present on epimera II. Regarding leg chaetotaxy the species belongs to the first group in Jack's nomenclature.

DISCUSSION
G. estherae n. sp. is parasite on T. mauritanica in Malta. In the past, it could have been confused with G. loricata. Among the species found on Tarentola sp., several species with ventral scale-like setae have been described. These species are distributed from Egypt to the Canary Islands (Fig. 4). Is it possible that some species have been misidentified? Certainly: Girot (1969) has questioned the identities of the two mites parasitic on Tarentola mauritanica: he noted a possible misidentification by the French naturalist Goux (1947) that described the development of "G. loricata", the venter of which being covered by scale-like setae. Girot contested Goux's identification: this mite originated from insular population studied from small islands around Sicilia (Isola de Ustica) and was fixed itself on the reptile's legs and not on the ventral scales (contrarily to G. loricata), and differed notably by the ventral setae, the number of inter-epimeral setae.
Since 1969, the identity of Girot's material, the Moroccan species of Tarentola, was enriched by subspecies descriptions (Geniez et al. 1999). Evidently, fifty years later, we can suspect that the parasites of the genus Tarentola were poorly investigated in regard of their putative diversity on the host complex: three species have been described from the West- Some other geckobians exhibit scale like setae as well: they were found on Gekkonidae, all from the Old World. Their presence coud be explained by overlapping niches of the hosts and contamination.
Since the Mesozoic, the modern geographic distribution of the phyllodactylid's parasites has seen different waves of colonization by the Phyllodactylidae, the rhythm being given by the different Mediterranean crisis. In the state of the present knowledge, we know that (Carranza et al. 2000 2. These species are classed into the Jack's first group defined by leg chaetotaxy. 3. The coxal setae of these species are homogenous in shape, stout and brush like (epimera 2, 3, 4).
4. The solenidiotaxy is identical on the tarsi I and II.

Eyes are present.
Because the host Tarentola mauritanica is now subdivided in several endemics and vicariants, we should consider realistic the drift of the fragmented host populations that determined analogous drift of the parasitic organisms. The early colonization of the New World by Tarentola's lineages could explain the absence of "New World geckobians with scale-like setae", because the African native Phyllodactylids migrated too recently to expand the lineage of the parasites, a current character of insular colonization.
In conclusion, it is highly probable that during each journey, the African Phyllodactylids have carried and disseminated the parasites, and that the parasites speciated to survive whereas the hosts differentiated. In the Mediterranean region, populations of T. mauritanica were examined including the molecular approach: a new species and subspecies were described from Tunisia (Joger and Bshaenia, 2010). The analysis of the parasitic mites bore by each of these populations should update the specific parasites for each host, species or subspecies The analysis should confirm or not if G. latastei (the second common species, widely distributed in Mediterranean Region) was submitted to the same process. Geckobia latastei and G. loricata are common in Morocco, Spain (Gil-Collado et al., 1985), France, Italy (Berlese, 1892), and Western Mediterranean Islands (Corsica, Balearic Islands...). If we try a synthesis of the knowledge, we know that: a) The genus Geckobia Mégnin, 1878 is specialized on the Gekkotans, everywhere the Geckkoes are present, and the main rule regarding hosts seems to be the geckobian's stenoxeny. Briefly, two cases can be observed, that are conform to the current knowledge (Boots et al. 2009). (i) when some hosts extended their distribution as the invasive geckoes did, the endemic native geckoes have kept their parasites (Martinez Rivera et al. 2003, Bertrand et al. 2008. (e.i. the genus Hemidactylus: Bertrand andIneich 1989, Bochkov andMironov 2000;Domrow 1983;Martinez Rivera et al. 2003).
(ii) On the endemic hosts, long duration of host-parasite interactions permitted hyperspecialized mites with allometric deformations (body wider than long, strengthening of posterior legs...). Examples: the endemic species or relict populations on Nactus Kluge 1983; or Bavaya Roux 1913... (Bertrand 1987, Bertrand et al. 2008). The apomorphies denote a common origin, and are rooted in the past, or contrarily, are they solely due to independent evolution i.e. convergence? The scale like setae are more differentiated than simple ones, is the differentiation evidently prior to the host divergence? b) Which are the more helpful criteria for reconstruction of the geckobian history? The body shape, the scutum (well delimited or not), the ocular plates (size, connected to the scutum or free).... help to recognize and identify the species with ease. The leg chaetotaxy (number and shape of setae) is the more integrative criteria and has been privileged because chaetotaxic analysis had given hope to help in a comprehensive ordination of the species: Jack (1964) established several "groups" of species, and the chaetotaxy opened the door to phylogenetic hypothesis based on general axioms elaborated through François Grandjean's principles (Jack, 1964). However, this classification is descriptive and naturally limited by the few setae on each of article, the constancy of this number on genua. The Jack's groups are useful, but they cannot be taken as absolute evidence of phylogenetic relationships, because identical pattern may result from different processes (Bertrand et al. 1995). Contrarily, when some apomorphic characters are shared by several species belonging to the same chaetotaxic group, a common origin may be suspected, and if the hosts are phylogenetically close, the radiation becomes highly probable.
Concluding remarks: pending questions and paths for future research on the Mediterranean geckobians, the genus Tarentola, and the Gekkonidae a) The Mediterranean host Tarentola spp.: The geckoes are differentiated since a long time as it is attested by the fossil gekkotans identified during the Mesozoic (Cretaceous: Borsuk-Białynicka, 1990, Arnold andPoinar, 2008) that is revealed by the distribution Gamble et al. 2008b. Even if the scale mites have indeed switched from insects to vertebrates more than 200 My ago, as speculated by Bochkov and O'Connor (2006), the Reptilian Mediterranean hosts were already available. Recent studies showed that in Mediterranean Region, the Gekkotans diverged at least since Miocene crisis because of the fragmentations of the distribution. Consecutively, the comprehension of the taxonomy of the genus Tarentola was improved by identification of some vicariant subspecies (Moroccan, Tunisian...) (Carranza et al. 2002, Geniez et al. 1999, Joger and Bshaenia, 2010. So, an updating "which hosts with which parasite" becomes more and more suitable now. b) The "two" historic parasites of Tarentola spp.: two species latastei and loricata were described on this host around the Mediterranean Basin and they can parasite simultaneously on the same individuals of this wall-gecko, one being preferably fixed between the toes, and the second on other favorable parts of the host's body. On the other hand, since the Berlese's description (1892), we know that the intimate interactions of the highly specialized by ectoparasitic mites obliged the parasite to a strict stenoxeny (Hirst, 1926, Bochkov andMironov, 2000). This observation of two species sharing the same individual popularized the thought that some species are heteromorphic (Hirst, 1926). More recently, Girot (1969) demonstrated that two true species can indeed share T. mauritanica: two geckobians of different lineages, one of them with ventral scale like setae, have succeeded, independently, in adaptation to the same host. Note that whereas geckobians colonized the five continents, and even if the host Tarentola is also present in the Americas (Jamaica, Cuba...), none of the species described in the New World possesses "scale-like ventral setae" (orally confirmed by Ricardo Paredes Leon, pers. com.). Contrarily, several species around the Mediterranean Basin exhibit such ventral setae.
c) The Mediterranean geckobians with scale like setae As the Maltese new species, several Mediterranean geckobians exhibit a central part of the ventral surface that is covered by scale-like setae. These special shaped setae have been described notably on G. loricata Berlese, 1892 in the Western Basin of the Mediterranean Sea, and on G. tarentulae Trägårdh, 1905, parasite of the Egyptian Tarentola annularis (Geoffroy-de-Saint-Hilaire, 1827).
Species with these special ventral setae have been found also on the Canary Islands: Geckobia tinerfensis and G. canariensis, parasites of Tarentola delalandei Steindachner, 1891; T. boettgeri Joger and Bischoff, 1983and T. gomeriensis Joger and Bischoff, 1983(Zapatero-Ramos et al., 1989. Other geckobians with such setae were identified in Asian, and around the Mediterranean Region (Bertrand et al., 1999, Bertrand et al. in press): these species are not parasitic on Phyllodactylidae but on Gekkonidae.
These data underline that new directions for researches with Pterygosomatids are now opened. This way is parallel to the development of the more modern data on the host populations (Carranza et al. 2002, Carranza andArnold, 2006) . Further research could offer the opportunities to test the possible common of Eastern and Western geckobians with ventral scale setae.