Loriciferans, cycliophorans and micrognathozoans are amongst the latest groups of animals to be discovered. Other than all being microscopic, they have very different body plans and are not closely related. Loriciferans were originally assigned to the Aschelminthes. However, both new molecular and ultrastructural researches have shown that Aschelminthes consist of two unrelated groups, Cycloneuralia and Gnathifera. Cycloneuralia may be included in the Ecdysozoa, including all molting invertebrates, and Gnathifera are more closely related to Platyhelminthes. The phylum Loricifera shares many apomorphic characters (e.g., scalids on the introvert) with both Priapulida and Kinorhyncha, and can be included in the taxon Scalidophora, a subgroup of Cycloneuralia. Cycliophora was originally allied to the Entoprocta and Ectoprocta (Bryozoa) based on ultrastructual research. Subsequent molecular data show they may be related to Rotifera and Acanthocephala, within the taxon Gnathifera. The phylogenetic position of Cycliophora is therefore not settled, and more ultrastructural and molecular data are needed. Micrognathozoa is the most recent major group of animals to be described. They show strong affinities with both Rotifera and Gnathostomulida (within the taxon Gnathifera), especially in the fine structure of the pharyngeal apparatus, where the jaw elements have cuticular rods with osmiophilic cores. Furthermore the micrognathozoans have two rows of multiciliated cells that form a locomotory organ, similar to that seen in some gastrotrichs and interstitial annelids. This character is never seen in Rotifera or in the monociliated Gnathostomulida. Rotifera and Acanthocephala always have a syncytial epidermis (Syndermata). Micrognathozoa lack this characteristic feature. Therefore, they are postulated to be placed basally in the Gnathifera, either as a sister-group to Gnathostomulida or as a sister-group to Rotifera + Acanthocephala.
Since the discovery of the Loricifera (Kristensen, 1983) 11 species have been described (Higgins and Kristensen, 1986; Kristensen and Shirayama, 1988; Todaro and Kristensen, 1998), and more than 80 recorded species await description (Kristensen, 1991a, b; Gad, 2000). The loriciferans seem to be one of the most abundant groups of meiofauna in the deep-sea, but also occur in mud on shallower water (Soetaert et al., 1984), and they may be one of the dominant meiofauna groups. Recently, Warwick (2000) has postulated that loriciferans are simply paedomorphic priapulids. This was based on the fact that the larvae of some loriciferans (Higgins and Kristensen, 1986) look superficially similar to the larvae of some priapulids (van der Land, 1970; Higgins et al., 1993). No comments were made concerning the unique adult features of loriciferans, including the cuticular buccal tube, myoepithelial pharyngeal apparatus with placoids (Kristensen, 1991a, b) and the very complex life cycle of the different species of loriciferans (Kristensen, 1999). Below I will discuss the different types of life cycles in loriciferans, totally different from all other known meiobenthic animals.
Cycliophora is the most recently described phylum, consisting of marine acoelomate epibionts on setae of mouthlimbs of lobsters (Funch and Kristensen, 1995). Today three species are known, all in the genus Symbion. The only described species, Symbion pandora, has a complicated life history with several attached and several free-living stages. The life cycle varies between asexual and sexual reproduction (Funch and Kristensen, 1997). The asexual and the sexual parts of the life cycle each have a unique larva. The life cycle seems to be more complex than we originally thought (Funch and Kristensen, 1995), and accordingly the life cycle has been redrawn in this paper. Their phylogenetic position is still uncertain. Cycliophorans have been considered to be a sister-group to Entoprocta and Ectoprocta (Funch and Kristensen, 1995; Funch, 1996), a view mainly based on ultrastructural investigation, or alternatively as a sister-group to the clade Rotifera-Acanthocephala (Winnepenninckx et al., 1998), a conclusion arrived at on the basis of molecular evidence.
Micrognathozoa was described by Kristensen and Funch (2000) as a new class within the Gnathifera (see Ahlrichs, 1995, 1997). This clade now contains Gnathostomulida, Micrognathozoa and Syndermata. Syndermata consists of all Rotifera (Eurotatoria and Seisonida) and the parasitic Acanthocephala. The sister-group relationship between Gnathostomulida and Rotifera + Acanthocephala was first mentioned by Rieger and Tyler (1995). The discovery of Micrognathozoa, in a cold spring from Greenland, strongly supports the above mentioned relationship (Ahlrichs, 1997; Herlyn and Ehlers, 1997). Only one species, Limnognathia maerski, has been described.
CURRENT STATE OF KNOWLEDGE
Adult loriciferans are bilaterally symmetrical marine metazoans 108–485 μm long (Higgins and Kristensen, 1986). The body is divided into five regions: mouth cone, head (introvert), neck, thorax and abdomen. The introvert of the adult has nine rings of scalids, which may have both sensory and locomotory functions. Two families, Nanaloricidae and Pliciloricidae, have been described, and two new families are under description. Nanaloricidae with one genus (Nanaloricus), is acoelomate both as larva and adult. The abdomen of the adult may be covered with a thick cuticular lorica consisting of 6 plates. Pliciloricidae with the two genera (Pliciloricus and Rugiloricus) may be acoelomate as larva, and pseudocoelomate as adult (Kristensen, 1995). The abdomen is usually covered with a thin lorica with many folds (plicae, in Higgins and Kristensen, 1986). The lorica of pliciloricids seems always to lack thick cuticular plates. All investigated loriciferans have chitin in the procuticle (Neuhaus et al., 1997).
The first loriciferans had already been collected in the seventies, but the first species, Nanaloricus mysticus, was described by me in 1983, from shell gravel off the coast of Brittany, France (Kristensen, 1983). The reason why it took so long from discovery to description was that these microscopic animals have a very complex life cycle, with a larva (Higgins-larva) which looks totally different from the adults (Kristensen, 1991a). Both the Higgins larvae and the adults may adhere strongly to sand grains or gravel by adhesive glands, so a new technique, freshwater-shocking on large samples of sand to release the animals from the bottom sediments had to be developed. Eleven species have been described, but about 80 other undescribed species have been collected, and many of the new species are found in deep-sea sediments (Gad, 2000). All known species are marine, microscopic, and live within the bottom sediments as adults. The Higgins-larvae can have very large swimming-toes (flippers) and have been taken free in the plankton. Alternatively, the larvae may be benthic and have tube-shaped toes with adhesive glands. Usually there exists a sexual secondary dimorphism in the first ring (clavoscalids) and last ring (trichoscalids) of scalids.
In the family Nanaloricidae the male has the middorsal, dorsolateral and ventrolateral pairs of clavoscalids divided into three branches but the two midventral scalids are undivided, so the clavoscalids consist of 20 chemoreceptors; the female always has only eight primary clavoscalids (Kristensen, 1991a). Sexual dimorphism may also be apparent in the trichoscalids (Kristensen, 1991b). In the male of Nanaloricus the two lateroventral secondary elements are modified as claw-shaped organs, called claspers. The claspers are assumed to function in copulation. Recently sexual dimorphism has also been recognised in the trichoscalids of Rugiloricus, where the female has seven of the 15 trichoscalids reduced in length (Kristensen, unpublished).
In several species of Rugiloricus the larva develops a mature ovary (neoteny), and this produces between 4 to 12 larvae inside the old larval exuvium (Kristensen and Brooke, 2001). These neotenous larvae look exactly like the larva emerging from the free-laid egg. The Rugiloricus-larvae may be common for most of the year, and outnumber the adults by 100:1. Postlarvae, when they are present (seemingly absent in Pliciloricus), lack toes, setae and locomotory spines on the abdomen. Often the postlarva appears very similar to the adults, but it always has fewer scalids and a thinner lorical cuticle than the males and females, and it lacks the gonads. The life cycle of Rugiloricus is surprisingly variable (Fig. 1). It varies from one species to another and, more confusingly, may vary depending on the feeding condition of the larva in the same species.
Three types of life cycles have been discovered in Rugiloricus. In R. carolinensis the Higgins larva moults direct to a mature male or female (Fig. 1A, B, C and H1, H2). In R. cauliculus the Higgins larva moults into a postlarva without toes and gonads. The postlarva moults into the adult (Fig. 1A, B, C, G and H1, H2). In a new species of Rugiloricus from Faroe Bank (North Atlantic Sea) the Higgins larva (Fig. 1D) grows bigger than the adult and develops a large single ovary. This larva forms a cyst-like stage (Fig. 1E). Inside the cyst moulting occurs and, after forming several eggs from the ovary, the Higgins larva totally disintegrates. The eggs mature to neotenous larvae inside the cyst walls. The larvae escape later from the cyst (Fig. 1F). The neotenous larva moults several times and grows to a new very large larva, which again develops an ovary. This part of the life cycle, called the parthenogenic life cycle (Fig. 1D, E, and F), recur and so explains why the Higgins larvae can outnumber the adults by a factor of 100. Under certain unspecified conditions the Higgins larva moults to a postlarva, which again moults into the adult. It is unclear if other species of Rugiloricus also have this parthenogenic life cycle, but there is some evidence that both R. carolinensis and R. cauliculus may also form neotenous larvae (Higgins and Kristensen, 1986; Kristensen, 1999; Kristensen and Brooke, 2001). Recently Gad (2000) has found neotenous larvae in several species of deep-sea pliciloricid loriciferans.
The fine structure of the mouth cone, the introvert with scalids, and the circumenteric brain indicate that this meiofaunal phylum is monophyletic, related to Kinorhyncha and Priapulida. The three groups are referred to as the Scalidophora (see Lemburg, 1995; Schmidt-Rhaesa et al., 1998). Autapomorphies for Loricifera may include the unique location in both males and females of the protonephridia inside the gonads, the complexity of the spinoscalids with muscles, and the special toe glands. The hexagonal arrangement in both adults and larvae of the internal armature in the buccal apparatus may be plesiomorphic, a shared character with Nematoda and Nematomorpha (Nematoida, subgroup of Cycloneuralia). Annulation of the flexible buccal tube, telescoping of the mouth cone and three rows of placoids in the triradiated pharynx bulb are found only in Tardigrada and Loricifera, but these similarities between tardigrades and loriciferans may be either convergent or plesiomorphic to all Cycloneuralia and Tardigrada (Kristensen, 2002). The thick lorica consisting of plates seems to be developed only in the family Nanaloricidae, for which it is a convincing autapomorphic character. The plesiomorphic condition in both adults and larvae of the lorica seems to be a thin lorica with many plicae.
Unfortunately, no molecular data are yet available for loriciferans.
The marine Cycliophora is the most recently described phylum in the animal kingdom. Only one microscopical species, Symbion pandora, has been described (Funch and Kristensen, 1995). The sessile stage had been known since the late sixties, but was never described (Funch and Kristensen, 1997). It is approximately 350 μm in length and 100 μm wide, and is attached with an acellular stalk and adhesive disc to crustacean mouth parts, which can be encrusted with thousands of individuals of the sessile stage. Investigations of mouth appendages of many decapod crustaceans world-wide have resulted in further discoveries of this new type of epibiont, with a very complex life cycle, living mainly on the setae of lobsters' mouthparts.
The sessile stage is the feeding stage, and has a circular mouth surrounded by a ring of compound cilia used for filtering small food particles, such as bacteria or food particles from the host. Cycliophorans are acoelomate metazoans with bilateral symmetry and a well-differentiated cuticle, and the epidermis is cellular rather than syncytial. They cannot, therefore, be assigned to the Syndermata (Rotifera and Acanthocephala). The internal anatomy is relatively simple, with a U-shaped gut and the anus located outside the ciliated feeding apparatus, a similar arrangement to that found in Ectoprocta (anus outside the tentacular crown, Nielsen, 2001) and some sessile rotifers (anus outside the corona, Winnepenninckx et al., 1998).
It is the life cycle consisting of an asexual and a sexual generation (metagenesis) of the cycliophorans that makes the phylum unique (Fig. 2). The asexual part of the life cycle consists of two types of feeding animals. The first type houses the so-called Pandora larva, the second type houses the primary male and female inside a brooding chamber. Furthermore, the sessile animal shows internal budding, whereby it loses the head (buccal funnel) and entire digestive system, replacing them with a new bud arising from embryonic cells in the posterior part of the trunk. This process of internal budding is repeated several times in the life of the sessile animal and is coupled to the asexual formation of the Pandora larva. The Pandora larva forms a new feeding-stage, also by internal budding. It settles very close to the maternal individual, and in this way the epibiont can multiply very quickly and completely cover the mouth limbs of its host. Later, the feeding stages may switch to produce a primary male or a sexually mature female instead of the Pandora larva. This marks the beginning of the sexual part of the life cycle. The primary male is released from the brooding chamber without a penis or a true testis. It settles on a feeding animal. Only if there is a young female inside this feeding animal can the attached primary male develop one or two new secondary males inside by internal asexual budding. Each secondary male has a tubular cuticular penis and a single compartment with spermatozoa. The female is about double the size of the primary male, and she looks similar to the Pandora larva except for lacking internal budding; rather anteriorly she has a single large oocyte.
Fertilisation of the egg occurs inside the female, perhaps just before or during the release of the female from the brooding chamber. She settles very close to her maternal feeding animal, and broods inside herself a new type of larva, the chordoid larva, so named after a mesodermal chordoid structure of approximate 50 stacked muscle cells (see Funch, 1996). The growing larva engulfs all her tissue, leaving only the cuticle. The chordoid larva hatches from the female exuvium, head first. It is a good swimmer and represents the dispersal stage. This larva settles on a new host and by internal budding forms a new feeding animal, so a new life cycle can begin. The complex life cycle of Symbion pandora as illustrated in Figure 2 differs from the original description only by the recognition of the free swimming secondary male (Fig. 2.17), where it can be shown that it develops by budding inside the primary male (Kristensen and Funch, 2001). We had not originally discovered the free-swimming secondary male because of its tiny size, only 37 μm long, so easily passing through our 63 μm net.
The type species was found on the Norway lobster, Nephrops norvegicus from Scandinavian waters. More recently two other and still undescribed species of the cycliophorans have been found on the mouth appendages of other lobsters such as Homarus gammarus (Europe) and H. americanus (North America). So far the cycliophorans have only been found on marine decapods from the Northern hemisphere. Records of cycliophorans from other crustaceans and nematodes all seem to be false alarms. All these records seem to represent chonotrich ciliates. Only by transmission electron microscopy it can be resolved that the ciliate only consists of one cell with several nuclei. Even so, the convergences between cycliophorans, and especially the chonotrich, Isochonopsis kergueleni, are surprising. In particular, this ciliate has a cryptogemmic budding like the cycliophorans (for detail, see Batisse and Crumeyrolle, 1988). Therefore the chonotrich ciliates often are recorded as cycliophorans on copepods or nematodes.
In the original description we stated that the Cycliophora are related to Entoprocta/Ectoprocta (Funch and Kristensen, 1995). The overall similarities with sessile rotifers (Collothecacea), and the presence of a dwarf male with the penial structure in both groups were considered to be superficial resemblances. This reluctance to ally cycliophorans and rotifers was based mainly on the ultrastructure of the epidermis covered with a cuticle and the lack of the mastax. All known rotifers have a mastax and a syncytial epidermis (Sørensen et al., 2000). These authors placed cycliophorans close to entoprocts on the basis of a careful cladistic approach. Unfortunately the phylogenetic relationships were only examined in the light of morphological evidence. The molecular data (study of 18S rRNA/DNA sequences) indicate that cycliophorans are a sister-group to the Syndermata which includes Rotifera and Acanthocephala (Winnepenninckx et al., 1998). Most recently Giribet et al. (2000) placed the cycliophorans within the Gnathifera. Their investigation was a combined approach based on both 18S rDNA sequences and morphology across the range of triploblasts, but their placement of Cycliophora was entirely dependent on our molecular data (1,542 base pairs of the 18S rRNA gene sequences) from alcohol-preserved material of Symbion pandora (see Winnepenninckx et al., 1998). The phylogenetic position of Cycliophora is, therefore, very uncertain, and new ultrastructural and molecular studies are urgently needed.
A new microscopic animal, Limnognathia maerski, with complicated jaws, was found in a cold spring at Disko Island, Greenland in 1994, but first described by Kristensen and Funch (2000), where it was assigned to a new class, Micrognathozoa, in the Gnathifera (Ahlrichs, 1997). The clade Gnathifera now includes Gnathostomulida, Micrognathozoa and Syndermata (Eurotataria, Acanthocephala and Seisonida), see Figure 3.
Limnognathia maerski has dorsal plates formed by an intracellular matrix, as in Rotifera and Acanthocephala (Storch and Welsch, 1969), but ventrally the epidermis is “naked” only with a thin glycocalyx. The gross anatomy of the body consists of a two-part head, accordion-like thorax and ovoid abdomen with a small tail. The total length of the holotype was only 142 μm, and the maximum width of the abdomen was 55 μm. As in many marine interstitial animals, e.g., gnathostomulids, gastrotrichs and polychaetes, special forms of tactile bristles or sensoria are found on the body. The locomotory organ consists of two rows of cells with cilia. These cells are similar to the ciliophores in annelids (Eibye-Jacobsen and Kristensen, 1994; Kristensen and Niilonen, 1982). An adhesive ciliary pad is located ventro-posteriorly. It consists of ten ciliated cells, which form a very sticky glue. The cells do not contain the two types of granulae which are characteristic in the duo-gland system (Fig. 16, Kristensen and Niilonen, 1982). Instead the ten cells are dominated by the cross-striated ciliary root passing through the entire cell. The cilia are ordered in register like those of the trunk ciliophores. The ciliary pad is very different from the adhesive toes of rotifers, gastrotrichs and annelids, and the structure may be a good autapomorphy for the micrognathozoans.
Because the new taxon lacks syncytia, a key character for the Rotifera and Acanthocephala (Syndermata, Ahlrichs, 1997), the Micrognathozoa cannot be included in Syndermata. Two pairs of protonephridia are found in the trunk of Limnognathia maerski, the nephridiopore of which seems to open laterally on the abdomen. The terminal cells are monociliated, a plesiomorphic character Micrognathozoa shares with Gnathostomulida (Ax, 1996). Only the female reproductive system has been observed, suggesting that Limnognathia maerski reproduces by parthenogenesis, as do many other freshwater meiofaunal animals, such as gastrotrichs, rotifers and heterotardigrades. The reproductive system is anatomically simple, and it seems that the two ovaries derive nutrition directly from the midgut, a feature also seen in freshwater chaetonotoid gastrotrichs. There is a possibility that the female lays two types of eggs, as in limnic gastrotrichs and rotifers: a smooth egg may be quick-developing summer egg, and a sculptured equivalent may be the resting winter stage (Funch and Kristensen, 2002).
Most of the digestive system in micrognathozoans is anatomically very simple, except for the pharyngeal apparatus with large array of true cuticular elements, the sclerites. In fact, Limnognathia maerski has one of the most complicated jaw systems in microscopical animals, and it is still not well understood. Until now only the main parts have been described by SEM investigations (Kristensen and Funch, 2000). They consist of two ventral basal plates, each plates associated with large teeth (manus), two dorsal oral lamellae and two lateral pharyngeal lamellae, two fibularia with small teeth anteriorly, and three sets of jaws, where the main jaws (mid jaws) are fused and consist of a dentarium (basket formed), and an articularium (Sørensen, 2002). The posterior part of the main jaws has a paired cauda, as in some species of gnathostomulids (Riedl and Rieger, 1972). Furthermore, most of the lumen of the pharynx is covered with a thin cuticle, and several accessory sclerites attaching the ventral and dorsal jaws are still not fully described (Sørensen, 2001). All sclerites seem to be built of microtubules, consisting of lucent, osmiophobic material surrounding a dense osmiophilic core. This structure was first described by Rieger and Tyler (1995) in the jaws of Gnathostomulida and Rotifera. The rest of the digestive system is surprisingly simple. A short dorsal esophagus leads into the large midgut, which occupies most of the thoracic region and the anterior part of the abdomen. The midgut lacks cilia completely, and has only a brush border of short microvilli. A true cuticle-lined hindgut is lacking. The anus is located dorsally and opens only periodically, as seen in Gnathostomulida and some Gastrostricha that also lack a permanent anus (Knauss, 1979).
The ultrastructure of the micrognathozoan jaws is similar to jaws of advanced scleroperalian gnathostomulids (Rieger and Tyler, 1995; Herlyn and Ehlers, 1997). However, the jaw apparatus might also possess some of the characteristics of the rotiferan mastax (Koehler and Hayes, 1969). Jaw-like structures are found in other protostome taxa as well, for instance in the proboscises of kalyptorhynch turbellarians, in dorvilleid polychaetes and aplacophoran molluscs, but studies of their ultrastructure (e.g., polychaetes, Purschke, 1987) show that none of these jaws is homologous with the jaws in Gnathostomulida, Rotifera, and Micrognathozoa (Sørensen, 2000; Sørensen et al., 2000).
Our proposed phylogeny of the clade Gnathifera is shown in Figure 3. The solid line is the suggestion of Kristensen and Funch (2000). The dotted line with a question mark denotes the suggestion of Ahlrichs (1997) that Acanthocephala and Seisonida are sister groups. The broken line with a question mark is our alternative suggestion (Kristensen and Funch, 1995) that Gnathostomulida and Micrognathozoa are sister groups. This preliminarily proposed idea was later confirmed by molecular data (Garey, personal communication). The black squares with numbers 1–11 mark synapomorphies: 1. Gnathifera. Cuticular jaws with tubes composed of lucent material surrounding an electron dense core. 2. Gnathostomulida. Monociliated epidermis. Hermaphrodites with complex reproductive system. 3. Micrognathozoa + Syndermata. Epidermis with apical intracellular matrix and dorsal and lateral trunk regions without motile cilia. 4. Filospermoidea. Elongation of the body with a long and slender rostrum. Compact jaws with 8 to 12 longitudinally arranged rods in lamellae symphyses. 5. Bursovaginoidea. Fibularization of the jaw apophyses. Complex female organs with a bursa and often a vagina. Male opening with an injectory penis often consisting of 8 to 12 intracellular rod-like structure. Spermatozoon strongly modified with lack of flagellar or centriolar structures. 6. Micrognathozoa. Ventral locomotory organ consists of two rows of multiciliated cells (trunk ciliophores). Ventral epidermis covered only with a thin glycocalyx. Cuticular outer oral plate and dorsal and lateral intracellulary plates present. The solid parts of the jaws very complex with paired basal plates, each associated with denticulate manus, paired striated oral and pharyngeal lamellae, paired fibularia and three sets of paired jaws. 7. Syndermata. No motile trunk ciliation. Eutelic, syncytial epidermis with apical crypts. Spermatozoon with anterior inserted flagellum. 8. Eurotatoria + Acanthocephala. Spermatozoon without acrosome. 9. Seisonida. Epizoic. The head ciliation consists of few tufts of stiff bristles. The trophi forms an anterior cuticular sheath. Presence of spermatophore. 10. Eurotatoria. Corona, unpaired retrocerebral glands, female reproductive organs with vitellarium. 11. Acanthocephala. Endoparasitic with a two-host life cycle. Epidermal lacunar system. Proboscis with intraepidermal hooks. Females with an uterine bell.
Molecular data are still not available for Micrognathozoa, but a study of 18S rRNA/DNA sequences is underway to be published (Garey, personal communication).
Loriciferans were originally assigned to the paraphyletic or polyphyletic group called Aschelminthes (Kristensen, 1983). Subsequently, both molecular (Winnepenninckx et al., 1995) and ultrastructural (Ehlers et al., 1996; Ahlrichs, 1997) studies have shown that Aschelminthes consist of two unrelated groups, Cycloneuralia and Gnathifera. Cycloneuralia may belong within the Ecdysozoa, including all molting invertebrates (Aguinaldo et al., 1997; Garey and Schmidt-Rhaesa, 1998; Schmidt-Rhaesa et al., 1998), while the Gnathifera are more closely related to either Platyhelminthes or Annelida.
The phylum Loricifera clearly shares many apomorphic characters (e.g., scalids on the introvert) with both Priapulida and Kinorhyncha, and can be included in the taxon Scalidophora (Ehlers et al., 1996). Furthermore they have several plesiomorphic characters, such as a myoepitheal pharyngeal bulb, comparable to the other Cycloneuralia, especially the Nematoda and Gastrotricha. The parasitic Nematomorpha, whose adults lack the pharyngeal bulb, has a larva with an introvert similar to Scalidophora. The loriciferans have, therefore, also been assigned to the Cephalorhyncha that includes Priapulida, Kinorhyncha and Nematomorpha (Malakhov, 1980; Adrianov et al., 1990). The phylogenetic position remains controversial, but the close relationship of loriciferans to kinorhynchs and priapulids can hardly be doubted (Ehlers et al., 1996; Neuhaus et al., 1996; Nielsen, 2001).
The idea that loriciferans are no more than “neotenous” or progenetic priapulids (Warwick, 2000) is based on misunderstandings of the supposed resemblances between adult loriciferans and the loricate larvae of priapulids. There are very few similarities, and it is incorrect that they are “virtually identical in morphology, except for the absence of gonads” (Warwick, 2000). It is correct that there are some similarities between the larvae of priapulids and loriciferans, but nearly all these similarities are plesiomorphic characters seen in all Cephalorhyncha (Nematoda, Nematomorpha, Priapulida, Kinorhyncha and Loricifera, see Nielsen, 2001). Furthermore if loriciferans are evolved from priapulids by progenesis, the larvae in loriciferans should not be present. This is the case in many progenetic polychaetes, e.g.,Neotenotrocha sterreri (see Eibye-Jacobsen and Kristensen, 1994; Westheide, 1987), which secondarily are adapted to the interstitial way of life. Loriciferans do not lack larvae. In fact, they have several types of larva; ironically one type produces a mature ovary (i.e., neoteny). Therefore, I reject the hypothesis that the loriciferans are simply progenetic priapulids. Loricifera is a monophyletic group, and is the sister group to the Priapulida + Kinorhyncha clade, within the Scalidophora (Nielsen, 2001).
In our original description (Funch and Kristensen, 1995) the Cycliophora was related to the Entoprocta and Ectoprocta (Bryozoa) on the basis of ultrastructual research. Cladistic morphological analyses (Zrzavý et al., 1998; Sørensen et al., 2000) place the Cycliophora as sister-group to Entoprocta. On the other hand, Cavalier-Smith (1998) considered them to be members of the phylum Kamptozoa (=Entoprocta). This latter hypothesis should be rejected because it necessitates a total redefinition of the Entoprocta. As the name implies, entoprocts always have the anus inside the tentacular crown (Nielsen, 2001), whereas the cycliophorans have the anus outside the ciliated mouth ring (Funch and Kristensen, 1997). However, our molecular data (Winnepenninckx et al., 1998) show they may be related to Rotifera + Acanthocephala inside the taxon Gnathifera (Giribet et al., 2000). The phylogenetic position of Cycliophora is, therefore, not settled, and more ultrastructural and molecular research is needed.
Micrognathozoa is the latest described group of the new animals. They show strong affinities with both Rotifera and Gnathostomulida within the taxon Gnathifera, especially in the fine structure of the pharyngeal apparatus. Only in these three groups do the jaw elements have cuticular rods with osmiophilic cores. Such microtubular structures are found in the fulcrum of all Rotifera and in several cuticular sclerites of both Gnathostomulida and Micrognathozoa (Sørensen, 2000). Furthermore, the micrognathozoans have two rows of multiciliated cells that forms a locomotory organ, similar to that in some gastrotrichs and interstitial annelids. This character is never seen in Rotifera, nor in the monociliated Gnathostomulida. Rotifera and Acanthocephala always have syncytia in the epidermis (Syndermata). Micrognathozoa lack this characteristic feature and, therefore, are postulated to be more basal in the Gnathifera, as a sister-group either to Gnathostomulida or Rotifera/Acanthocephala.
The jaws of Gnathostomulida, Rotifera, and Micrognathozoa, with their microtubes, are considered to be homologous structures. Limnognathia maerski is probably related to both Rotifera and Gnathostomulida. Syndermata + Micrognathozoa might be sister-groups, while Gnathostomulida could be the sistergroup to this assemblage (Kristensen and Funch, 2000). Gnathostomulida was originally described as an order of the Platyhelminthes (Ax, 1956). More recently Ax (1996) postulated a sister-group relationship of Gnathostomulida with Platyhelminthes (“Plathelminthomorpha hypothesis”). This hypothesis is rejected (Kristensen and Funch, 2000), and the finding of Limnognathia maerski supports the monophyly of Gnathifera (see Ahlrichs, 1997). The ultrastructural similarities of the new animal to certain chaetonotoid gastrotrichs and interstitial polychaetes may be either convergent or plesiomorphic features.
The discovery of Micrognathozoa has clearly thrown new light on the polyphyletic origin of the “so-called” Pseudocoelomata or Aschelminthes. Gnathifera are more closely related to other platyzoan acoelomates (Gastrotricha and Platyhelminthes) than to introvertan pseudocoelomates. Cycliophora may be included in Platyzoa, too (Giribet et al., 2000). Introverta (Cycloneuralia without Gastrotricha) form part of the clade Ecdysozoa (Aguinaldo et al., 1997). Already in 1991—before the molecular evidence for Ecdysozoa—I had stated (Kristensen, 1991a) that some aschelminth groups (later named Scalidophora) and arthropods may be related. Therefore, similarities (i.e., chitin in procuticle) between Panarthropoda and Scalidophora could be true homologies. Proarthropoda (Tardigrada and Onychophora) could have developed from Scalidophora-like animals similar to Opabinia and Anomalocaris of the Cambrian Burgess Shale Fauna, and not from annelid-like animals without a moulting cuticle. The evidence for a clade Ecdysozoa comprising all moulting invertebrates, is strongly supported by the discovery of Loricifera, especially the life cycles of the new pliciloricid loriciferans that indicate a close relationship between Panarthropoda and Scalidophora.
From the Symposium Lesser-Known Protostome Taxa: Evolution, Development, and Ecology presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 3–7 January 2001, at Chicago, Illinois.
I gratefully acknowledge the collaboration with my teacher Dr. Robert P. Higgins for the 20 yr of work together on meiobenthos, and to Dr. James Garey, who invited me to SICB 2001 in Chicago and for the interesting discussions on a number of matters relevant to this article. Especially I thank Dr. Andreas Schmidt-Rhaesa, Dr. Simon Conway Morris and an anonymous reviewer who carefully read the manuscript and added helpful comments and linguistic corrections.
I am indebted to Mrs. Stine B. Elle and Mrs. Birgitte Rubæk who made the final drawings of loriciferans, cycliophorans and micrognathozoans. Danish financial assistance came from the Carlsberg Foundation (970345/30-488) and the Danish Natural Science Research Council (9701589).