Revision of the family Carabodidae (Acari, Oribatida) VII. Redefinition of the genus Malgasodes; redescription of M. curvisetus Mahunka, 2000; and complementary description of M. hungarorum Mahunka, 2010. Phylogenetic relationships between Malgasodes, Bovicarabodes, Afticarabodes, Congocepheus and Cavaecarabodes are discussed

Abstract The genus Malgasodes is redefined; the type species M. curvisetus Mahunka, 2000, is redescribed by means of studies using optic and Scanning Electron Microsopy (SEM), and a complementary description of M. hungarorum Mahunka, 2000 is included. Comparison of genera Malgasodes Mahunka, 2000, Bovicarabodes Fernandez, Theron, Rollard, 2013a, Cavaecarabodes Fernandez, Theron, Rollard, Rodriguez Castillo, 2014, Afticarabodes Fernandez, Theron, Rollard, 2013b, and Congocepheus Balogh, 1958 is made. Problems concerning chaetotaxy, regressive evolution and neotrichy are explained and phylogenetic relationships between Malgasodes, Bovicarabodes, Afticarabodes, Congocepheus and Cavaecarabodes are discussed.


Introduction
For several years we have been working on a revision of the family Carabodidae. Many years ago, upon commencing our work on this family, we studied large collections of material, principally from Madagascar, Gabon, Namibia, South Africa and Argentina; later, other collections were studied with material from Antilles, Vietnam, Central African Republic, Morocco, Congo, Thailand, Korea, China, Costa Rica, Brazil, Paraguay, etc. We rapidly understood how essential it was to study type material deposited in different Museums; and in general collaboration was very good, but in some instances type material was not available on loan (National Natural History Museum, Budapest), leading to problems in the development of our study. A significant quantity of unavailable material is from Madagascar. We were fortunate as more than 4000 specimens from Madagascar were obtained from the Betsch and Paulian expedition team (Fernandez and Cleva 2010) and we managed to find specimens of all species unavailable on loan from Madagascar within these samples.
In this paper, we report on Malgasodes curvisetus (from the Madagascar collection at the MNHN) and Malgasodes hungarorum made available on loan from the Natural history Museum of Geneva, thus the totality of the species of this genus, with optic microscopy and SEM.
Identification of M. curvisetus was easy, due to several characteristics pointed out in the original description by Mahunka (2000), such as shape of interlamellar, lamellar, rostral, and notogastral setae; shape of sensillus and elevated interlamellar process.

Materials and methods
Specimens studied with light microscopy were macerated in lactic acid and observed in the same medium using the open-mount technique (cavity slide and cover slip) described by Grandjean (1949) and Krantz and Walter (2009). Drawings were made using an Olympus BHC compound microscope (Olympus, Rungis, France) equipped with a drawing tube. To aid observations, some specimens were stained with chlorazol black E (Coineau 1974). Specimens studied with the aid of scanning electron microscopy (SEM) were prepared as follows: specimens preserved in ethanol were carefully rinsed by sucking them into a Pasteur pipette several times, then transferring them for 2 hours to buffered glutaraldehyde (2.5%) in Sörensen phosphate buffer (pH 7.4; 0.1 m). After postfixation for 2 hours in buffered 2% OsO4 solution and rinsing in buffered solution, all specimens were dehydrated in a series of graded ethanols and dried in a critical point apparatus. Specimens were mounted on Al-stubs with double-sided sticky tape, after which they were gold-coated in a sputter apparatus (Alberti and Fernandez 1988;Fernandez 1990a, 1990b;Alberti et al. 1991;Alberti et al. 1997;Alberti et al. 2007).
Measurements taken: total length (tip of rostrum to posterior edge of notogaster) and width (widest part of notogaster) in micrometres (μm). The description of leg chaetotaxy (SEM, standard, polarized and phase contrast microscopes) of Malgasodes curvisetus should be considered provisory as detailed study of leg chaetotaxy during ontogenetic development is necessary. Setal formulae of the legs include the number of solenidia (in parentheses); tarsal setal formulae include the famulus (ε).
Lyrifissures difficult to observe; three pairs present, first pair situated at level of h 3 setae ( Figure 12), another situated anterior to p 3 setae and ips situated between p 1 and p 2 . Lyrifissures not visible in SEM, only under optic observation (Figures 12,14).
Lateral region. Tutorium (tu) clearly visible as a strongly curving cuticular thickening. Between lamellae and tutorium a deep supratutorial depression (s.tu.d) running parallel to both structures (Figures 6,7,12,19). Bothridia cup-shaped with smooth bothridial ring (bo. ri); bo.ri incomplete with bothridial tooth (bo.to) clearly visible (Figures 8,12,15,19); sensillus (si) uncinate with small barbs, curving upward ( Pedotectum I, prominent extended lamina. Pedotectum II, small polyhedral lamina, rounded edges. Humeral apophysis more or less triangular; basally slightly convex, immediately becoming concave; posterior bothridial zone overlapping anterior tip (Figure12); aligned irregular furrows (a.i.f) delimited by rod-like cuticular structures (Figures 6, 19, indicated by arrow), crossing h.ap. SEM observations made from two different lateral angles (Figures 6, 19) in order to clarify the relative position, shape and disposition of different prodorsal and notogastral structures and setae. In order to show positioning of legs during "legs folding process' specimens are shown with legs in place and alternatively with distended legs. Only one lyrifissure visible at level of h 3 setae Discidium not discernible. Only one depression, situated behind leg IV, used conceal tibia and tarsus (paraxial side) ( Figure 6 indicated by X) during leg folding process (see Fernandez et al. 2013b). unique pairs visible in posterior view. Lyrifissure ips can be uniquely identified on account of placement between setae p 1 and p 2 ; the other, probably ih, situated between p 3 and h 3 .
In this position the zone between ventral and notogastral plates is easily discernible, separated due to particular tectal border (indicated in Figure 14 by ·).
Ventral region. Subcapitulum with three pairs of setae (a, m, h) clearly visible; insertion zone m and a setae simple, smooth (Figures 16, 18). Epimera hardly discernible; only bo.sj clearly visible as shallow furrow crossing the medial plane ( Figure 16).
Genital plate more or less similar in size to anal plate ( Figure 5); anal plate small, sharply tipped ( Figure 16); paraxial border of anal plates with small teeth ( Legs. Legs presenting all characteristics observed in other Carabodidae, but lateral setae of genua are particular ( Figure 11). The (u) pair of all tarsi are without barbs but rugous, particular ( Figure 9).

Remarks.
Anterior part of n.a.d extending to d.sj, together with notogastral setae situated in this zone preventing clear observation. The e.i.p. is an elevated process situated not far from d.sj. The posterior part of e.i.p descending steeply at d.sj level. We consider p.p.d to be absent.
Two pairs of setae situated close to d.sj are probably c 1 ,and c 2 setae, but without immature stases it was impossible to establish with certainty.
Border of tectum in Malgasodes is unlike all others; in that it is not prolonged by a limbus, and that a space exists between the ventral plate and the notogaster.
Various inclinations in lateral view are provided to permit better understanding of the disposition of notogastral setae and the steeply descending posterior part of e.i.p towards d.sj.

Malgasodes hungarorum Mahunka, 2010 Figures 24-27
Remarks. Only one bleached specimen was available for study. The condition of this material was poor, for this reason a decision was made to add only characteristics and figures considered inadequate in the original description by Mahunka (2000). The ventral zone (text and figure) contained adequate information; some omitted aspects will be addressed in text. In contrast the dorsal and lateral descriptions and figures require redescription and new figures.
Fourteen pairs of notogastral setae, two pairs outside posterior notogastral anterior depression directing forward; four pairs inside notogastral anterior depression, three pairs near dorso-sejugal furrow, fourth pair distant; four pairs marginally to notogastral anterior depression; four pairs marginally to notogaster; two pairs lyrifissures present. Humeral apophysis easily discernible.
Posterior prodorsal depression (p.p.d) absent (Figures 24, 25). Setae in, medium length, situated on e.i.p, similar to long simple notogastral setae situated adjacent to and inside n.a.d, in all cases in setae directing backward and entangled with the cited notogastral setae (Figures 24, 25 ).
Fourteen pairs of notogastral setae, two pairs situated posterior to n.a.d, one far from posterior border of n.a.d; both pairs directing forward, not exceeding d.sj; four pairs inside n.a.d, three pairs situated near d.sj extending backward or laterally (indicated in Figure 25 with X); fourth pair situated far from d.sj, extending forward or laterally, not exceeding d.sj (indicated in Figure 25 withJ); four pairs situated marginally to n.a.d, extending forward; four pairs, situated marginally to notogaster, directing backward, these four pairs are possibly named as h 3 , p 1 , p 2 , p 3 ( Figure 25).
Lyrifissures difficult to observe; two pairs present, first pair situated at level of h 3 setae, the other situated between p 3 and p 2 (Figure 24).
Remarkable border tectum, not prolonged by a limbus, space existing between the notogastral and ventral plates (Figures 24, indicated by arrows ·).
Only two pairs of lyrifissures visible at level of h 3 setae and between p 3 and p 2 . Discidium not discernible. Sejugal zone depressed. Semicircular ridges at level of acetabulum IV (Figure 24). Ventral region. Well described by Mahunka (2000: page 89 and Figure 6) only adding: the a.g.d is clearly visible.

Introduction
Many years of study and previous publications lead us to the point where we are able to compare these genera. Several new collections with large numbers of specimens have become available, but it will take several years to describe this new material. A decision was made to first complete the comparison of the cited genera before commencing further descriptions.
In previous publications, we described Bovicarabodes, Afticarabodes, Cavaecarabodes, and studied and redescribed most of the species of Congocepheus. All genera present several characteristics in common such as: notogastral anterior depression, posterior prodorsal depression, superior cornea of naso, elevated interlamellar process, notogastral setae situated around the notogastral anterior depression; humeral apophysis; anterior genital furrow. We observed two evolutive phenomena: regression and neotrichy in Cavaecarabodes anouchkae and Congocepheus germani, respectively.
Malgasodes presents particular characteristics and some characters in common with genera mentioned above. Firstly we explain the situation of Malgasodes, after which we complete the comparison.
We studied specimens of both M. hungarorum (paratype) on loan from MHNG and M. curviseta unavailable from the Hungarian Natural History Museum, Budapest, but fortunately found in the Madagascar collection of Professor J-M Betsch (between 1968 and 1973), deposited in Museum National d'Histoire Naturelle, Paris, France (MNHN).
In the original description of Malgasodes, Mahunka (2000: 87) indicated: "Notogaster with a wide deep, roughly semicircular, hollow anteriorly, well framed laterally and posteriorly". "Fourteen pairs of notogastral setae: ten pairs long medially directed, recurved and four pairs short, phylliform, in marginal position. Of these 14 pairs, eight pairs arise around hollow, two pairs arise posteriorly and remaing four pairs present in posteromarginal position".
However, in Figure 1 (page 88) it is clear that four pairs of the eight indicated arise from and originate inside the n.a.d ("hollow"sensu Mahunka) and are not all situated around the n.a.d. In our studies we confirmed that four pairs are inside n.a.d (see redescription of M. cuvisetus and M. hungarorum).
In the comparison with other genera Mahunka indicated "Its allies are in the genre Congocepheus Balogh, 1958 andBaloghodes Mahunka, 1986. However, Congocepheus is clearly distinguished from Malgasodes by the form of the ventral structure, mainly the ventral plate and the position of the lyrifissures iad. Species of the genus Baloghodes are characterized by the absence of the anteromedian hollow and covered lateral margin".
Mahunka only considerered the ventral characteristics in Congocepheus but there are other more important characteristics (see Table 2).
The comparison with Baloghodes is hard to understand as Baloghodes differs greately from Malgasodes.

The posterior prodorsal depression
There are often small differences between immature stases of oribatids during ontogeny, and these differences obey precise rules, for example the number of genital setae is fewer in protonymph than in deuteronymph, which is fewer than in the tritonymph (3 pairs, 4 pairs and 5 pairs), but adult stases differ greatly from immature stases. Studies were not found on ontogenetic development in groups of Carabodidae where the posterior prodorsal depression is present in adult stases, but some exist for Carabodes (Grandjean 1949, Andre 1975, Bellido 1978, Ermilov 2011, Reeves 1992 and Yoshiobodes (Reeves 1997).
In 1949 Grandjean provided information on orthogenetic development of Carabodes labyrinthicus, while immature stases of the same species were studied by Andre in more detail in 1975. Evidently immature stases are very different to adults. Immature stases of Carabodidae are non-sclerotized with soft and fragile bodies. We searched through previous studies to determine if some characteristics could be found in the groups under study.
Species in these studies relate to genera without posterior prodorsal depressions (p.p.d), or notogastral anterior depressions (n.a.d) in adult stases, but there are very interesting structures on the immature prodorsum of C. willmanni Bernini, 1975 and Y. irmayi (Balogh & Mahunka, 1969) (Reeves 1997). In C. willmanni Bernini, 1975(Bellido 1978) the prodorsum presents a foveate sclerite situated between lamellar and interlamelar setae ( Figure 6, 7 in Bellido op.cit); this structure can be observed in the larva, but in the protonynph the sclerite is clearly differentiated as a bowl-like structure, and in the deutonymph this zone acquires coloration (considerered melanisation), and sclerotization (Bellido 1978: 423). Reeves (1997), indicated (page 320) "The scalloped edged depression on the prodorsum of protonymphs, deutonymphs and tritonymphs appears similar to the foveate sclerite found in immature described by Bellido (1978) of Carabodes willmanni Bernini." In other cases, such as Carabodes subarticus (Ermilov 2011 -in russian) the figures of immatures are very deformed, and the presence of the sclerotized field between interlamellar and lamellar setae is only indicated in the description of the deutonymph. On figures 3.1 and 4.1, a structure exists between the interlamellar setae and posterior to the lamellar setae, probably the structure referred to by Bellido and Reeves. The descriptions of Andre (1975) and Reeves (1992) do not show this structure in immature stases. Bellido (1978: 424) indicated that the existence of this prodorsal microsclerite permits easy differentiation between C. willmanni and C. labyrinthicus.
We were only able to study adults, not having immatures available, but we consider that the posterior prodorsal depression is probably the same depression found on the immature stases of Carabodes and Yoshiobodes, and in adult stases on Congocepheus, Bovicarabodes, Cavaecarabodes and Afticarabodes.
Another important element is that the prodorsal structure of the prodorsal zone in Yoshiodes was not discussed by the author in text (Figure 9 dorsosejugal region, page 321), similar to Carabodes interruptus (Fig. 26) and Carabodes pentasetosus (Fig. 34) (in Reeves 1992) where a depression, probably a prodorsal posterior depression of reduced size, exists in the medial region of the prodorsum of the adult, but for these two species no reports on ontogenetic stages exist.

The anterior notogastral depression
All genera cited in this comparison present an anterior notogastral depression and a particular relation between this depression and notogastral setae.
In genera Bovicarabodes, Afticarabodes, Cavaecarabodes and Congocepheus notogastral setae are related to the anterior notogastral depression, but setae are never found inside it; in the case of Malgasodes, the situation is very different and four pairs of setae are observed inside the n.a.d.
The presence of the notogastral anterior depression and the setal disposition around and/or inside it, involves a very interesting problem overlooked in descriptions of several genera of the family Carabodidae.
In Malgasodes, the situation is very different. The notogastral anterior depression is present, but the distribution of notogastral setae is unlike all others, with a particular distribution: four pairs of marginal setae p 1 , p 2 , p 3 and h 3 in the normal position; four pairs marginal to and two pairs posterior to n.a.d; finally four pairs inside n.a.d, of which two pairs are close to d.sj, these are probably c 1 and c 2 setae. Lyrifissure numbers are reduced.
The disposition of the notogastral setae is very particular, demonstrating displacement of most of the setae directing to the n.a.d (Figure 34), and making accurate notation of the setae near impossible.
We have not been able to advance further, as ontogenetic studies are required; but this displacement process is found near the dorsosejugal furrow (dis or das) and this may indicate a very interesting evolutionary problem related to segmentation of the opisthosoma, the segmentary origin of the prodorsum and with terminology of the body in general.
Another very interesting condition is the reduction process found in Caveacarabodes anouchkae (Figure 31) where there are only twelve pairs of setae, with reduction of c 1 and c 2 setae. The area of setal reduction is indicated in Figure 33.
A completely different situation occurs between the genera Congocepheus and Malgasodes.
In Congocepheus the setae involved in neotrichy are easily identified by a "neotrichous territory" that follows the border of n.a.d. However, in the case of Malgasodes, a reduction process has occurred where two or three lyrifissures have disappeared. A particular process exists where the setae are directed towards the n.a.d (Figure 34), and a large zone of the notogaster is therefore void of setae due to setal migration. Four pairs of setae migrate to inside the n.a.d, with two pairs extending to near the d.sj furrow; four pairs are found near the border of n.a.d; two pairs situated slightly posterior to n.a.d and four pairs are situated in the posterior marginal zone of notogaster. Despite no modification in the number of setae (always orthotrichy bideficience) the migration process impedes identification of the setae involved in the process; most probably the two pairs situated near d.sj furrow are c 1 and c 2 , but it is impossible to affirm. The displacement of setae does not involve all of them, because setae p 1 , p 2 , p 3 and h 3 , situated on the marginal zone, are not affected by this process. More complex still, the setae involved are probably from two different segments. The migration process found in Malgasodes is very remarkable and necessitates further ontogenetic study.