Systematic revision of the adeleid haemogregarines, with creation of Bartazoon n. g., reassignment of Hepatozoon argantis Garnham, 1954 to Hemolivia, and molecular data on Hemolivia stellata

Life cycles and molecular data for terrestrial haemogregarines are reviewed in this article. Collection material was re-examined: Hepatozoon argantis Garnham, 1954 in Argas brumpti was reassigned to Hemolivia as Hemolivia argantis (Garnham, 1954) n. comb.; parasite DNA was extracted from a tick crush on smear of an archived slide of Hemolivia stellata in Amblyomma rotondatum, then the 18S ssrRNA gene was amplified by PCR. A systematic revision of the group is proposed, based on biological life cycles and phylogenetic reconstruction. Four types of life cycles, based on parasite vector, vertebrate host and the characteristics of their development, are defined. We propose combining species, based on their biology, into four groups (types I, II, III and IV). The characters of each type are defined and associated with a type genus and a type species. The biological characters of each type are associated with a different genus and a type species. The phylogenetic reconstruction with sequences deposited in the databases and our own new sequence of Hemolivia stellata is consistent with this classification. The classification is as follows: Type I, Hepatozoon Miller, 1908, type species H. perniciosum Miller, 1908; Type II, Karyolysus Labbé, 1894, type species K. lacertae (Danilewsky, 1886) Reichenow, 1913; Type III Hemolivia Petit et al., 1990, type species H. stellata, Petit et al., 1990; and Type IV: Bartazoon n. g., type species B. breinli (Mackerras, 1960).


Introduction
The haemogregarines form a group of particularly diverse heteroxenous adeleid coccidia parasites which have exploited all environments, terrestrial or aquatic, and become adapted to numerous vertebrate hosts, i.e. chelonians, crocodiles and other reptiles, amphibians, fishes and many mammals.
(i) In the aquatic environment -the transmission of parasites is obligatorily achieved either by predation between vertebrates [38,70] or through vectors in close contact with the vertebrate hosts.
For example, in the wild, an Eimeria of fish may be transmitted from fish to fish by cannibalism or via a paratenic host such as a shrimp [38,70] but not, in natural conditions, by shedding oocysts in the water where they would be immediately diluted.
The haemogregarines of aquatic hosts are transmitted by leeches or by arthropods in which the sexual part of the cycle develops. The sporogony of Haemogregarina stepanowi Danilewsky, 1885 [14] develops in the leech which transmits the infection when feeding on the turtle [60]. The oocysts in the leech are asporate and produce free sporozoites which are inoculated to the turtle. In some vectors of the haemogregarines of fish, a further stage develops from the sporogony: a merogony, in the leech for Cyrilia Lainson, 1981 [36], in the isopod for Desseria Siddall, 1995 [65]; the vertebrate host would become infected when ingesting the vector.
(ii) In the terrestrial environment -the life cycle of haemogregarines comprises roughly four stages: merogony and gamogony in the vertebrate host, and fertilisation and sporogony in the invertebrate. Merogony in the vector is absent.
In addition to the classical cycle in which transmission is achieved by the bite of the vector or its ingestion by the vertebrate host, a second mode of transmission was acquired by some species: transmission by predation between vertebrates [39]. This mode of transmission is shared by all species when the alimentary diet of the host does not include the direct ingestion of the vector by the vertebrate host.
When, for example, the vector is a mosquito and the vertebrate host a snake, it is obvious that a haemogregarine cannot be transmitted regularly by ingestion of the mosquito. There must be a second vertebrate host which eats insects, develops cysts in its tissues and is part of the diet of the snake. This second host may or may not develop, in addition to cysts, the entire cycle of the parasite.
Transmission by predation is characterised for all parasites by (i) the absence of specificity; (ii) a wide repartition of infective stages (here the cyst) in the organism of the host. These principles, established for nematodes, are valid for the haemogregarines producing cysts which are disseminated in numerous organs [7].
When the vector is a mite or a tick, the sporogonic development may follow two courses: (i) in one step: the sporogony evolves directly from zygote to oocyst, sporoblasts and sporocysts inside the same envelop, like in Hepatozoon perniciosum Miller, 1908 [51] or (ii) in two steps: oocysts undergo the first division to produce motile sporokinetes instead of sporoblasts. Sporokinetes, after the rupture of the oocyst envelop, invade new cells of the host; sporokinetes complete their development into sporoblasts and sporocysts either in the same host, like in Hemolivia stellata Petit et al., 1990 [58] or in the next host generation when they invade the oocytes of the mite, like in Karyolysus Labbé, 1894 [35].
In both instances, the vertebrate host ingests either directly the invertebrate host or cysts from the tissues of another vertebrate host. It was suggested that sporocysts of Hemolivia mauritanica or H. stellata Petit et al., 1990 [58] might also be excreted with the faeces of the tick and be infective to susceptible hosts or transported by paratenic hosts.
As pointed out by Smith (1996) [68], a great many haemogregarines were described on the basis of gametocyte morphology and very often designated as Hepatozoon spp. or Haemogregarina spp. However, only the observation of stages in the vector may indicate the generic position of the parasite [71]. Through the years, the nomenclature has evolved, while new life cycles have been unravelled. For example, Hemogregarina mauritanica Sergent and Sergent, 1904 [64] studied by Laveran (1905) [41] and Brumpt (1938) [5] was renamed successively Hepatozoon mauritanicum by Michel (1973) [50] and later Hemolivia mauritanica by Landau and Paperna (1997) [40].
The genus Hepatozoon, well defined by the morphology and the life cycle of its type species, H. perniciosum Miller, 1908, has over time become a heterogeneous group of species with diverse life cycles and which, according to work by Barta et al. (2012), is paraphyletic (see Discussion) [3].
The genus Hemolivia, which is clearly defined by its morphological and biological features, has recently been investigated by molecular biology and its phylogenetic relationship with Hepatozoon studied [3,27,34]. A molecular analysis of two of the three Hemolivia species, Hemolivia mauritanica from Testudo graeca and Testudo marginata [27,34], Hemolivia mariae from Egernia stokesii and Tiliqua rugosa [3,34], was performed, as well as for Hemolivia sp. from Rhinoclemmys Pulcherrima manni [34]. A number of haemogregarines of reptiles and amphibians could probably be assigned to the genus Hemolivia if their life cycle were known.
In this communication, we (i) present molecular data on H. stellata obtained from 25-year-old archived original material; (ii) partially re-describe the haemogregarine Hepatozoon argantis Garnham, 1954 [20] and reassign it to the genus Hemolivia; and (iii) analyse the haemogregarines' known life cycles and propose Bartazoon n. g.

Hemolivia argantis
In 1986, PCC Garnham deposited part of his collection of slides at the Welcome Trust, including many types of haemosporidians [21] and the rest of his collection at the Muséum National d'Histoire Naturelle (MNHN) in Paris, France. In the MNHN's collection, we found part of the original material used to describe Hepatozoon argantis Garnham, 1954 in the tick Argas brumpti Neumann, 1907 [53]. This material consists of sections of the tick with a massive infection by a haemogregarine which we identified as belonging to the genus Hemolivia and not Hepatozoon.
As we are certain that we are dealing with the original material and that only one species of haemogregarine is present, and because of the absence of type designation by the authors, this material should be considered as syntype. It consists of sections of the tick Argas brumpti on which the description of the haemogregarine H. argantis was based; the tick was identified as belonging to the original description by a photograph (Fig. 18) in Garnham 1954 [20], showing one of the sections of the tick and by drawings of oocysts of different stages. Some sections were stained by Ehrlich's haematoxylin and eosin, some by haemalun eosin, and smears of caecal contents by Giemsa stain.
A 25-year-old archived smear of crushed Amblyomma rotondatum (Koch, 1844) [31] which was part of the material used for the original description of H. stellata, containing many immature stellate oocysts (Fig. 1I), was used as a source of DNA for molecular characterisation of this parasite on the basis of 18S ssrRNA.
The coverslip was dismounted; the smear was scraped off the glass slide with a sterile scalpel; the material collected was incubated at 50°C in ATL buffer containing proteinase K until total digestion; DNA was extracted using the QIAamp DNA Mini Kit from Qiagen Ò following the manufacturer's recommendations. DNA was eluted in 50 lL of elution buffer and frozen at À30°C. DNA was amplified by a seminested PCR assay. The first amplification was performed with one pair of universal 18S rRNA oligonucleotide primers 2867 [5 0 -AACCTGGTTGATCCTGCCAG-3 0 ]/2868 [5 0 -TGATCCTTCTGCAGGTTCACCTAC-3 0 ], as described by Mathew et al. (2000) [47]. For the second step, two seminested PCRs were carried out with one external oligonucleotide primer of the first reaction paired with one internal Protozoan-specific primer derivate from Vilcins et al. [74] as follow: 2867/Hep900 [5 0 -CAAATCTAAGAATTTCACCTCT-GAC-3 0 ] and Hep300 [5 0 -GTTTCTGACCTATCAGCTTTC-GACG-3 0 ]/2868 amplifying to overlapping fragments of 939 bp and 1510 bp, respectively. The PCRs were run in a total volume of 20 lL, containing 1X High Fidelity PCR Buffer, 3 mM of MgSO 4 , 0.5 U of Platinum Ò Taq DNA Polymerase High Fidelity (Invitrogen TM ), 0.2 mM of each dNTP (Promega), 0.25 lM of each primer and 3 lL of original DNA template in the first reaction; 1 lL of the PCR product was used as a template in the second reaction. All PCRs were run on a Veriti Ò Thermal Cycler (Applied Biosystems Ò ). The PCR products were resolved by 1.5% agarose gel electrophoresis prior to purification by the QIAquick Ò PCR Purification Kit (Qiagen Ò ), following the manufacturer's recommendations. Purified products were eluted in 30 lL of nuclease-free H 2 O and frozen at À30°C. PCR products were prepared for sequencing in both directions, using the BigDye Ò Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems Ò ) and the respective oligonucleotide primers. The BigDye Ò reaction products were purified using the BigDye Ò XTerminator TM Purification Kit (Applied Biosystems Ò ), following the manufacturer's recommendations before being sequenced on a 3500xl Genetic Analyzer (Applied Biosystems Ò ).
Alignment and cross-checking of the sequences were performed with CLC Main Workbench 5.7 software (CLC bio) and a consensus sequence of 1816 bp was obtained by combining the two overlapping fragments. The nearly complete sequence (1816 bp) of the 18S ssrRNA gene recovered from the original archived material of Hemolivia stellata was deposited in GenBank under Accession Number KP881349. Our new sequence of H. stellata and 179 sequences of adeleid parasites (Table 2) were aligned using a Muscle algorithm [19]. Molecular phylogeny was performed by the Maximum Likelihood (ML) method with a GTR + C + I model, using PhyML 3.0 software [23]. Nodal robustness of the tree was evaluated by non-parametric bootstrapping (1000 replicates).

Results
Life cycles and morphology H. argantis was described in detail by Garnham, 1954 [20] and we will only provide a general outline: (i) in the haemocoelom: large immature oocysts with peripheral nuclei (Figs. 1A & 1B); (ii) round, immature oocysts with sporokinetes budding at the periphery of a large cytoplasmic mass ( Fig. 1C) or, apparently free in the oocyst (Fig. 1D); and (iii) sporocysts filled with sporozoites in the haemocoelom (Fig. 1F) and the gut contents of the tick.
In the re-examined material, we also found a few intracellular sporocysts beneath the intestinal epithelium and inside desquamated cells of the gut (Fig. 1E), suggesting that the initial site of development of sporokinetes into sporocysts is intracellular.
Hemolivia stellata was described by Petit et al., 1990 [58]. Schizonts of this species occur in erythrocytes and leucocytes of the cane toad Rhinella marina (Linnaeus, 1758) (as Bufo marinus) and gametocytes of this species occur in erythrocytes of this amphibian. Oocysts in the tick, Amblyomma rotondatum, are star shaped, and release sporokinetes that

Parasite sequences and phylogenetic analysis
The phylogenetic analysis included 166 published sequences of Hepatozoon (140), Karyolysus (8) and Hemolivia (18) parasites isolated from a variety of vertebrates, 13 other published sequences of Adeleidae as outgroups to root the tree and our sequence of H. stellata (Table 2).
A robust Hemolivia clade was obtained with H. stellata at the base (Fig. 2). It comprises all the Hemolivia sequences included in the phylogenetic construction but also a few sequences from Hepatozoon isolated in Australia from Varanus panoptes (Storr, 1980) and from Liasis fuscus Peters, 1873.
What was previously considered as the Hepatozoon group was paraphyletic. It was clearly divided into two major clades (Fig. 2).
The first clade was subdivided into two robust subclades containing respectively the Hepatozoon from carnivores and a mixed group containing some of the Hepatozoon of Mediterranean reptiles and the Karyolysus of the European lacertids.
The second clade was also subdivided, with the Hemolivia clade on the one hand and a group containing sequences of haemogregarines of reptiles and amphibians transmitted by biting insects (Diptera and Siphonaptera) on the other.
A sequence of a haemogregarine from the bird Oceanodroma melania (Bonaparte, 1854) was found at the base of this group, as well as a monophyletic group of sequences of parasites from Australian and South American marsupials, a monophyletic group of sequences from amphibian hosts from various geographic areas, a few sequences of haemogregarines from rodents and three sequences of parasites extracted from the liver of the bats Hipposideros cervinus (Gould, 1863) [59].

Discussion
Taxonomic status of H. argantis It appears clearly that the parasite in Argas does not belong to the genus Hepatozoon in which the sporoblasts mature into sporocysts inside the oocyst envelop. The sporogony evolves in two stages: (i) mature oocysts release motile sporokinetes, and (ii) sporokinetes invade the cells of the tick and develop into sporoblasts and sporocysts. Garnham in 1954 [20] did point to differences with the classical cycle of Hepatozoon and compared the parasite from Argas with another genus existing at the time, Karyolysus (Danilewskyi, 1886) [15], which also produces sporokinetes in a mite Lyponyssus. However, in the latter, sporokinetes invade the oocytes of the mite and mature only in the next generation. The authors choose to assign the haemogregarine in Argas to the genus Hepatozoon but noted that it might be a new genus. It is now clear that it belongs to the genus Hemolivia.

Host spectrum
The vectors, definitive hosts of terrestrial haemogregarines, fall into two groups: insects (Diptera and fleas) on the one hand, and haematophagous Acari (ticks and mites) on the other.
Mosquitoes and fleas are considered as hosts for Hepatozoon while Acari are vectors of the three existing genera: Hepatozoon, Karyolysus and Hemolivia.
Hepatozoon is at present a large gathering of species classified in this genus on the sole basis of the presence of gametocytes in the blood. Some species, previously considered as Hepatozoon, when their life cycles were unravelled, were assigned to the genus Hemolivia: H. mauritanica, H. mariae, H. stellata and here H. argantis. The development in the vector of many haemogregarines follows different courses which will be discussed in the next section.

Life cycles
The fundamental life cycle of a Coccidiomorpha consists of (i) the infective stage, i.e. the sporozoite, (ii) male and female gametes, and (iii) the zygote. Many adaptive additions to this simple scheme arose either to multiply the parasite in the vertebrate host (schizogony) or in the vector (sporogony) or to facilitate transmission according to the hosts and their life habits: free resistant stages ingested in the external environment, infective stages ingested by a paratenic host, bite by a vector and predation between vertebrate hosts.
The adeleids' fertilisation procedures, including syzygy, gametogenesis with maturation of a macrogamete and production of a small number of microgametes inside a common envelope, followed by the fertilisation of the macrogamete, were reported from several haemogregarines considered as Hepatozoon and Hemolivia.
Another mode of fertilisation is described in Hepatozoon of mammals transmitted by ticks or mites: Hepatozoon perniciosum Miller, 1908 [51] of the rat, the type species of the genus, H. canis, and H. americanum from Canidae: syngamy, which is the association of a pair of male and female gametes and their fusion without production of flagellate microgametes.
In the literature, syngamy was described in the gregarine Coelogregarina ephestiae Ghelelovitch, 1948 [22] from Ephestia kuehniella (Zeller, 1879). It would be a unique example of syngamy in the Gregarines and the question of their classification within the Gregarinomorpha or the Coccidiomorpha may be considered.

Molecular data of Hemolivia
The molecular data deposited under the name Hemolivia into the databanks reaches 21 sequences of the 18S ribosomal RNA gene: (i) The sequence JN211118 deposited by Barta et al. (2012) [3] as Hemolivia mariae Smallridge & Paperna, 1997 [67] was isolated from dried blood films containing gamonts from an experimentally infected Tiliqua rugosa Gray, 1825 from Australia. This sequence is not included into our tree (Fig. 2) because of its shortness and of poor overlap with the rest of the sequences. However, it clusters with the Hemolivia sequences that are long enough to overlap (data not shown). In Barta et al. (2012) [3], it also clusters with the sequence ''Hepatozoon'' sp. EU430236 that belongs to the Hemolivia clade in Kvičerová et al. (2014) [34] and in our analyses (Fig. 2).
(ii) The sequence HQ224961 also deposited by Barta et al. (2012) [3] as Hemolivia mariae Smallridge and Paperna, 1997 [67] is mistakenly referenced in GenBank while it is clearly stated in the text of the article that it is a sequence of Babesiosoma stableri Schmittner & McGhee, 1961 [62] obtained from Rana septentrionalis Baird, 1854 collected by hand from Lake Sasajewun, Algonquin Provincial Park, Ontario, Canada. As reported [3,34], this sequence also clusters in our analysis (Fig. 2) with the related Dactylosoma ranarum Labbé, 1894 [35].
(iv) The sequence KF270674 deposited as Hemolivia sp. by William et al. (2014) [78] was isolated from the blood of Panthera leo (Linnaeus, 1758) from Zambia. This sequence is not included in our tree (Fig. 2) because of its shortness (298 bp) and in addition appears more related to sequences of Adelina and Dactylosoma than to sequences of Hemolivia by BLAST [2].
(v) The 13 sequences KF992698 -KF992710 deposited as Hemolivia mauritanica by Kvičerová et al. (2014) [34] were isolated from Hyalomma aegyptium L., 1758 collected on T. graeca from Algeria, Iraq, Syria and Turkey, and on Testudo marginata Shoepf, 1789 from Greece. All these sequences cluster with the sequence KC512766 (Fig. 2) and seem to correspond to Hemolivia mauritanica and some of its variants. Finally, Karyolysus and Hemolivia have a similar life cycle in two stages but they differ in several respects, such as the vector (according to present knowledge): mite vs. tick; the vertebrate host: lizard vs. turtle; trans-ovarian transmission vs. direct transmission. Molecular biology shows the two genera to be separated into two distinct clades with a well bootstrap value for clade A (in which the sequences of Karyolysus are found) but not for clade B (in which the sequences of Hemolivia are found).

Phylogenetic analyses
Our analyses show the phylogenetic position of H. stellata to be at the base of a robust Hemolivia clade. This clade supports the monophyly of the genus Hemolivia, as previously reported [3,34], and includes several parasites assigned to the genus Hepatozoon: EU430231, EU430232 isolated from Varanus panoptes (Storr, 1980), EU430236 isolated from Liasis fuscus Peters, 1873 from Australia [76]. When more information on their life cycle is known, they might be reassigned to Hemolivia.
Siddall (1995) [65], Smith and Desser (1997) [69], analysing the morphology and life cycle traits of Hepatozoon species, thought that they consist of different genera. Barta 2014), using molecular tools came to the same conclusion [3,34]. Analyses also agree on the paraphyly of the genus Hepatozoon that contains the Hemolivia clade and appears clearly divided into two major clusters (Fig. 2). The first one is subdivided into two robust clades containing respectively the Hepatozoon from carnivores (v) the parasites of the ranids, grouping hosts from North America and Europe; and (vi) the parasites from Madagascan caecillid hosts; the rest forms a broad-range-host cluster containing parasites from diverse geographic areas, essentially isolated from reptiles and rodents, and from bats.
The nature of the haemogregarine described in the lizard Sphenodon punctatus and the associated tick Amblyomma sphenodonti in New Zealand remains unsolved. The lizard is considered as the only surviving member of the ancient reptilian order Sphenodontia and the parasites in the blood have an unusual morphology. According to the illustrations in Laird, 1950 [37], three types of gametocytes may be seen: (i) beanshaped intra-erythrocytic gametocytes of a classical type with a folded tail inside a surrounding envelope, (ii) intra-leucocytic smaller forms surrounded by an ellipsoid thick envelope which prevents full staining of the gametocyte (Fig. 3, in Laird 1950 [37], Fig. 1B, in Herbert et al. 2010 [29]), and (iii) small elongated parasites, with pointed ends (Figs. 1 and 2, in Laird 1950 [37], Fig. 1A in Herbert et al., 2010 [29] and Fig. 9 in Desser, 1978 [16]) considered by Laird to be merozoites. Herbert et al.
2010 described oocysts and sporocysts in the haemocoel of the tick. However, as there is no description of the initial stages (pairing and fertilisation) and the partial sequence (233 bp) is too small, the parasite's identity cannot be ascertained and the sequence was not included.
The majority of sequences in the phylogenetic tree fall into one of the four groups defined below and correspond to a different genus characterised by its cycle, hosts and vectors. However, a few sequences appear in a group which does not correspond to the life history supposed by their authors. The risk of errors, when classifying a haemogregarine according to molecular data only, is that, when indispensable information on their life cycle is lacking, many species with sequences deposited in the databanks are not identified correctly at the generic level.
The main sources of errors are: (i) when, in insect eating animals or their predators, sequences derived from tissues or organs where cysts from an undetermined haemogregarine are present. For example, the sequence derived from a Hepatozoon in the liver of a Chiroptera [59] should be interpreted with caution. We have observed on several occasions, cysts from a haemogregarine in sections of the liver from Miniopterus (Fig. 3) and considered them to be cysts from a Hepatozoon developing in a Dipteran ingested by the bat. Haemogregarine gametocytes were never found in the blood of any bat. These cysts are probably a dead end. (ii) When the host is polyparasitised and only one of the species is sequenced. (iii) When a vector from the wild is assumed to transmit the same parasite as the one seen or sequenced from the blood of the vertebrate host. The sequences of H. fitzimonsi KJ702453 and KR069084, originating from blood of a turtle, group with the Hepatozoon clade (=Bartazoon). The authors show images of sporocysts in the smears of ticks which may belong to ruptured oocysts of a Hepatozoon or of an accumulation of sporocysts from Hemolivia. The same applies for sequences EU43033 and EU43034 in samples extracted from ticks engorged on reptiles (Varanus panoptes, Liasis fuscus and Dendrolaphis pustulatus), from Australia. No description or data on the biology is attached to show that the parasite develops in the tick. The sequences extracted from the blood of Marsupials (EU430237 and EU430238) group with the Hepatozoon (=Bartazoon) clade. The authors found ticks on the host but did not demonstrate a role of the Acari in the transmission of the parasite.

Systematics
The type species of the genus Hepatozoon is H. perniciosum, which raises a real problem since most other species identified as Hepatozoon and Hemolivia or Karyolysus do not One stage Two stages Two stages One stage Table 2. List of the sequences used in the phylogenetic construction. The different columns give respectively the accession numbers of the sequences, the name of the parasites to which they are assigned, their vertebrate hosts, their isolation source, the country in which they have been found and the number of the associated reference in the references list. Unpub.: unpublished data only deposited in GenBank; a: experimentally fed on a naturally infected host; b: experimentally infected; c: collected from naturally infected wild host; d: Amblyrhynchus cristatus DNA detected in the last blood meal; n.a.: not available.  undergo the process of fertilisation by syngamy. We think that the mode of fertilisation is an important character and a fundamental part of the cycle of haemogregarines. It is at present associated with sporogony in a tick or a mite and schizogony and gametocytes in a mammal.
According to current knowledge on life cycles, terrestrial haemogregarines can be classified into four types: Type 1: Haemogregarines of mammals transmitted by ticks and mites; fertilisation by syngamy, sporogony in one stage, with complete sporocyst development inside the oocyst envelop: genus Hepatozoon Miller, 1908 [51], type species: Hepatozoon perniciosum Miller, 1908 [51].
Type 3: Haemogregarines of reptiles and amphibians transmitted by ticks; syzygy of gamonts followed by microgametogenesis with production of a small number of gametes inside a common envelop with the macrogamete; sporogony in two stages: mature oocysts release sporokinetes which penetrate inside new cells in which the sporocysts develop. No transovarian transmission occurs. Genus: Hemolivia Petit et al., 1990 [58], type species, H. stellata Petit et al., 1990 [58].
Surprisingly, Hemolivia and Karyolysus which were considered as biologically closely related belong in fact to two different clusters, the first one with the Bartazoon n. g. and the second one with the Hepatozoon of carnivores.

Conclusion
Classification, particularly of species into genera, aims at defining biological and morphological categories common to several species. We believe that creating a taxon for species or genera is much more helpful and less confusing than leaving parasites that are obviously different together. They may be easily synonymised if new elements are produced.
We propose (i) to reassign Hepatozoon argantis to the genus Hemolivia; (ii) the following new classification of terrestrial haemogregarines, consistent with the recent phylogenetic constructions: Haemogregarines of Type I: genus Hepatozoon; Haemogregarines of Type II: genus Karyolysus; Haemogregarines of Type III: genus Hemolivia; Haemogregarines of Type IV: genus Bartazoon.
This classification is consistent with current knowledge on biology and life cycles and with molecular data on species well identified by their life history.