Taxonomic revision of the Nippostrongylinae (Nematoda, Heligmonellidae) parasites of Muridae from the Australasian region. The genus Odilia Durette-Desset, 1973

The species of the genus Odilia Durette-Desset, 1973 (Heligmonellidae, Nippostrongylinae) are re-distributed among eight genera of which five are new. This classification is mainly based on certain characters of the synlophe not previously taken into account at the supraspecific level. These characters mainly include the presence or absence of a careen, the relative size of the ridges forming the careen, the development and position of ridge 1’, the development of the left ridge and right ridge, and the distribution of the largest ridges. Eighteen of the 20 known species are rearranged in the following genera: Odilia sensu stricto Durette-Desset, 1973 with Odilia mackerrasae (Mawson, 1961) as type species, Chisholmia n. gen. with Chisholmia bainae (Beveridge & Durette-Desset, 1992) n. comb. as type species, Equilophos n. gen. with Equilophos polyrhabdote (Mawson, 1961) n. comb. as type species, Hasegawanema n. gen. with Hasegawanema mamasaense (Hasegawa, Miyata & Syafruddin, 1999) n. comb. as type species, Hughjonestrongylus Digiani & Durette-Desset, 2014 with Hughjonestrongylus ennisae (Smales & Heinrich, 2010) as type species, Lesleyella n. gen. with Lesleyella wauensis (Smales, 2010) n. comb. as type and sole species, Parasabanema szalayi Smales & Heinrich, 2010, and Sanduanensis n. gen. with Sanduanensis dividua (Smales, 2010) as type and sole species. Odilia uromyos Mawson, 1961 and Odilia carinatae Smales, 2008 are not included in the new classification. A key to the proposed genera is provided. The new generic arrangement follows a distribution more related to the biogeographical areas than to the host groups.


Introduction
This paper contains a taxonomic revision of the Australasian Nippostrongylinae. In a preceding paper, we studied the Paraheligmonelloides complex, which was divided into four genera [4]. The present paper concerns the genus Odilia Durette-Desset, 1973, with 20 described species in murids from mainland Australia, Tasmania, Borneo, Sulawesi, and New Guinea.
The synlophe of these species (except O. uromyos) was described by Durette-Desset [5] and the genus Odilia (=Austrostrongylus sensu Durette-Desset, 1971, nec Chandler, 1924) was defined as follows: ''synlophe with axis of orientation directed from right-ventral line to left-dorsal line; hypertrophied lateral ridges; left-dorsal ridge almost as long as left-ventral ridge tending to the formation of a small careen, the latter remaining of moderate size. Gradient in ridge size lateromedian or ridges very numerous and of similar size'' [6].
After a gap of approximately 20 years, eight new species were described between 1992 and 2005: three from mainland Australia and Tasmania [1,11], and five from Indonesia [12,16]. Smales [20] proposed a key to the 15 species then known in the genus, mainly based on the number of cuticular ridges and, to a lesser extent, on the length of the gubernaculum. Since 2005, five other species from New Guinea have been described and assigned to Odilia [21][22][23]25], bringing the number of species in the genus to 20. At present, the composition of the genus is very heterogeneous, mainly due to the great variability of the synlophe among the species: some species possess a careen, whereas other species do not, the number of ridges varies from 14 to 35, the lateral ridges are not always well developed, and the ridge size also varies considerably.
It is likely, in one way, that the generic definition was sufficiently ambiguous to allow the inclusion of species with very different synlophes. In another way, however, some species do not even match the generic definition.
In an ongoing revision of the Heligmonellidae, it became necessary to review the specific composition of Odilia, and to attempt to group the species into new, possibly supraspecific taxa. This work is presented here, based mainly on the morphology of the synlophe, a complex structure involving numerous characters, some of which are proving to be taxonomically useful at the supraspecific level.

Materials and methods
The data were compiled from the published descriptions. The species whose synlophe could be analyzed were: Odilia mackerrasae (Mawson, 1961 [14], whose synlophes were not illustrated, were not included in the study. The methods used for the study and description of the synlophe follow the terms and criteria provided by Durette-Desset [8] and Durette-Desset & Digiani [9]. To indicate more accurately the position of the ridges around the body circumference, the body section may be divided primarily into hemispheres or sides: right side and left side determined by the sagittal axis (SA) (Fig. 1A), and dorsal side and ventral side determined by the frontal axis (passing through the lateral hypodermal cords) (FA) (Fig. 1B). Applying the same principle, the ridges situated on the dorsal side are named dorsal ridges, those on the left side left ridges, etc. More complex synlophes usually require for their description a division of the section into quadrants or even into octants. The intersection of the SA and FA determines four quadrants referred as to leftdorsal, right-dorsal, right-ventral, and left-ventral (Fig. 1C). Similarly, the intersection of the diagonals of the quadrants defined above determines four other quadrants which may be referred to as mid-dorsal, mid-right, mid-ventral, and mid-left (Fig. 1D). The ridges situated in the different quadrants are named following the same principle as right-ventral ridges, left-dorsal ridges, etc. (Figs. 1C and 1D). A division into octants results in sections identified as dorsal-right-dorsal, right-right-dorsal, right-right-ventral, ventral-right-ventral, ventral-left-ventral, left-left-ventral, left-left-dorsal, and dorsal-left-dorsal, and the ridges on them are named consequently as dorsal-right-dorsal ridges, right-right-dorsal ridges, etc. (Fig. 1E).
As proposed in a previous article [4], the terms ''right ridge'' and ''left ridge'' when used in the singular, indicate the single ridge closest to the right and left lateral fields, respectively.
In the Heligmosomoidea, the ridges are usually numbered according to an axis of orientation which separates them into two groups with tips pointing in opposing directions. The axis of orientation is always directed from the right-ventral quadrant to the left-dorsal quadrant and in the Nippostrongylinae, its inclination ranges from 1°to 90°on the sagittal axis, depending on the species. At 90°, it overlaps the frontal axis passing through the lateral cords. The numbering of the ridges always begins on the left side and by definition, ridge 1 is situated dorsally to the axis and ridge 1' is situated ventrally. If the axis is frontal, it separates one group of dorsal ridges numbered 1 to n and another of ventral ridges numbered 1' to n'. If the axis is oblique, it separates one group of right-dorsal ridges numbered 1 to n and another of left-ventral ridges numbered 1' to n' (Fig. 2A).
The left ridge may be homologous with ridge 1' ( Fig. 2A) or distinct from it (Figs. 2B-G). This seems to be a strong character which allowed the rearrangement of several species in the revision of genera such as in Neoheligmonella Durette-Desset, 1971 [3] and Paraheligmonelloides Fukumoto, Kamiya & Suzuki, 1980 [4].
The main synlophe characters used to separate the different genera were the following: (1) presence or absence of a careen; (2) size of ridges forming the careen; (3) position and development of ridge 1'; (4) development of the left ridge; (5) development of the right ridge; (6) size of the ridges; (7) distribution of the largest ridges; (8) presence of cuticular dilatations; and (9) discontinuity of ridges.
The description of the bursa follows Durette-Desset & Digiani [10]. Other characters, especially of the bursa and spicules are included in the generic definitions. The abbreviation SpL/BL refers to the spicule length as a proportion of the body length, expressed as a percentage. The nomenclature of the hosts and their taxonomy at the suprageneric level follow Musser & Carleton [18].

Discussion
Since some of the characters analyzed were frequently associated with one another, we were able to group the species treated into eight groups: Groups (1 and 2), species with a careen O. polyrhabdote (Fig. 2H) and O. similis (Fig. 3H).

Comments
Odilia uromyos is the most prevalent species present in Uromys spp. from Australia and Papua New Guinea [17,27]. However, its synlophe has never been described in transverse section of the body. It was described as having up to 40 small, apparently subequal ridges in the distal part of the male and 48 in the female [17]. This number of subequal ridges may correspond either to the genus Parasabanema or to Equilophos n. gen. However, other characters such as the bursal pattern, which is apparently 2-3 [17] and the ratio SpL/BL of 16% do not correspond with any of these two genera. Based on the available data, O. uromyos cannot be placed in the generic arrangement proposed above, although the species is validated by a number of characters including the number of ridges, bursal pattern, and spicule shape. It is temporarily considered as a Nippostrongylinae incertae sedis.
Odilia carinatae, also parasitic in Uromys spp., was described from Papua New Guinea [21]. The synlophe was described and illustrated in males and females but the orientation of the ridges is not completely clear and its interpretation is difficult. In both sexes, a careen is absent, the ridges are unequal in size, the mid-right ridges are among the largest ridges, and there is a left or left-dorsal cuticular dilatation. However, in the male the cuticular dilatation is more marked and the largest ridges on the left side are the left-dorsal ridges, apparently two dorsal and two ventral to the axis of orientation, whereas in the female, the largest left ridges are the mid-left ridges, apparently all ventral to the axis of orientation. The synlophe characters of the female evoke those of the genus Hughjonestrongylus, with several species reported from Melomys spp., Paramelomys, and Chiruromys but also Uromys spp., all from Papua New Guinea [4, 22, 24, 26, this work]. The synlophe characters in both sexes, as well as the spicular characters of the male [21], enable us to distinguish these specimens from all the other species treated herein: however, they cannot be placed in the generic arrangement proposed above. Therefore, it would be preferable to consider these specimens as Nippostrongylinae incertae sedis, while awaiting improved descriptions of the synlophe in both sexes.
Two other species were reported as Odilia sp. 1 and Odilia sp. 2, parasitic in Rattus cf. morotaiensis from the Molucca Islands, Indonesia [14]. The synlophes of both species were described but not illustrated, and the data provided are insufficient to assign the species to any of the genera proposed here. Table 1 provides a list of the species included in the former genus Odilia, with their new systematic position as proposed herein. Table 2 provides the list of genera of the ''Odilia'' complex, along with the species list, host spectrum, and biogeographical distribution.
At this point, it is interesting to note that all native rodents of Australia and New Guinea (Sahul region) belong to the Murinae [18]. Among them, two groups are recognized. The first and most speciose is a taxonomically diverse but phylogenetically connected group of old endemics [19]. Recent studies support the monophyly of the old endemics, the result of a single colonization event 5.1-5.5 million years ago and subsequent rapid diversification within the region [19]. This was followed by multiple dispersal events between Australia and New Guinea, related to sea level fluctuations [19].
The other group, called the new endemics, is composed of native species of Rattus, which are thought to be recent colonists (from about 1 million years ago) with a history independent from that of the remaining Sahulian murines [19].
The new generic arrangement proposed for the nippostrongylines belonging to the ''Odilia complex'' seems to follow a distribution which is apparently more related to different geographic areas than to definite host groups (see Table 2).
Species of Hughjonestrongylus, Sanduanensis, and Lesleyella are distributed among several genera of hosts belonging to different groups of the old endemics, but are all restricted to New Guinea.
Species of Parasabanema have a Sahulian distribution, being present in Australia and New Guinea and parasitizing members of different groups of the old endemics. Species of Equilophos also have a Sahulian distribution, with one species in an old endemic of New Guinea and another in a native Rattus of Australia.
Species of Odilia and Chisholmia are found in mainland Australia (one species in Tasmania), parasitizing members of different groups of old endemics but also species of native Rattus (new endemics).
All these taxa of Sahulian distribution are likely to have undergone high diversification accompanying that of their hosts, diversification which followed the first colonization of Sahul by the ancestors of the old endemics. The presence of certain genera (Equilophos, Parasabanema) in both Australia and New Guinea may also reflect the latter hosts' multiple dispersal events between these two areas. None of these genera with exclusively Sahulian distribution has species parasitic only in the new endemics. This may reinforce the hypothesis that they diversified mainly in the old endemics, whereas the presence of about four species in native species of Rattus could  (Table 2), with four out of five species found in endemic murines of Sulawesi [16], and only one species parasitic in West Papua (New Guinea) in one member of the old endemics [12]. Furthermore, species of Hasegawanema show certain synlophe characters, such as the left ridge being minute or replaced by a gap, which remind us strongly of species of Syafruddinema Digiani and Durette-Desset, 2014 [4], with two species parasitic in endemic murines of Sulawesi and a third in the Malay peninsula [4,16]. A similar close relationship with representatives of Sundaland (Southeast Asia and islands of the Sunda shelf) has also been found in other endemic heligmonellids of Sulawesi such as Hasanuddinia , with affinities with Rattustrongylus Ow-Yang et al., 1983 from peninsular Malaysia [13], or Maxomystrongylus Hasegawa & Syafruddin, 1997 which has congeners in Kalimantan (Borneo) [15,16]. Such a close relationship between the heligmonellids of Sulawesi and Sundaland is not unexpected since the endemic murid fauna of Sulawesi is most closely associated with that of the Asian mainland [16,18].
Key to the proposed genera