Fertilization in Wirenia argentea
According to the general presence of organs known to be involved in internal fertilization (seminal receptacles, copulatory stylets) and a sperm type characteristic for internal fertilization, most - if not all - Solenogastres species appear to deposit fertilized eggs . Reciprocal copulation is thought to be the general mechanism for sperm transfer [5, 25]. The low numbers of sperm in the seminal receptacles of our investigated adult specimens, however, raise an interesting notion. The gonad of sectioned specimens contained numerous ripening and mature sperm cells, but we did not observe any copulation behavior in the laboratory. Self-fertilization is known from various mollusks, most interestingly from certain hermaphroditic opisthobranch slugs with internal fertilization . Even if at this point there is no proof for self-fertilization in W. argentea, it is possible that the animals can switch from fertilization via copulation (indicated by presence of sperm in the seminal receptacles) to use of own sperm under stressful conditions, such as laboratory culture.
Oviposition in Solenogastres
In Wirenia argentea, eggs are deposited shortly after fertilization and prior to the first cleavage. Brood protection has been described for other species, such as Halomenia gravida, observed to carry larvae in the mantle cavity , and Epimenia babai, having at least the capacity of retaining eggs and larvae for several days in the mantle cavity . Wirenia argentea (present data), Nematomenia banyulensis , Neomenia carinata , and Rhopalomenia aglaophaeniae  deposit uncleaved embryos. Solenogasters lay their eggs in paired batches - correlating to the paired arrangement of spawning ducts - with several days to weeks between deposition events. Only in E. babai the eggs of each batch are enclosed in a thick mucus sheath , while generally there is only a thin mucus coating surrounding the egg hulls. The number of eggs deposited per batch varies largely between species and appears to be dependent on body size of the parent. In addition, more eggs are produced early in the breeding period [; personal observation]. Large E. babai specimens of 20 cm body length deposit up to 100 fertilized eggs at once and up to four such batches within one night . Thompson's single N. carinata specimen was 2 cm long and also produced up to 100 eggs per batch , while small species such as W. argentea deposit only 4-6 eggs per batch. Egg size, however, is not correlated in a linear way to average species size. The largest eggs with a diameter of 700 μm (including egg hull) have been found in N. carinata , a species with a maximum body size of 3 cm, while the eggs of the up to 20 cm long E. babai measure 350 μm , and the eggs of both N. banyulensis with a maximum body length of 3 cm  and W. argentea with a body length of about 0.5 cm are 200 μm in diameter. It has to be noted that in some cases (E. babai, probably also N. carinata) the space between embryo and egg hull is considerably larger than for example in W. argentea.
Hatching and larval behavior in Solenogastres
As described here for Wirenia argentea, solenogaster larvae usually hatch at a swimming stage when the prototroch ciliation is well developed [23, 28, 29]. In Epimenia australis, by contrast, the larvae leave the "fertilization membrane" (egg hull) much later, namely at the postlarval crawling stage . Since these observations are based on fixed material only, they require corroboration by in vivo studies. In contrast to the negative geotaxis observed in newly hatched larvae of Wirenia argentea, Neomenia carinata larvae seem to stay within a 2 cm range from the bottom . The basic information available for Epimenia babai [23, 28] shows similarities to our findings on W. argentea. In the former species, the younger larvae exhibit vivid spiral swimming, typical for spiralian trochophore-like larvae, and do not show any phototactic behavior. In E. babai, metamorphosis starts at days seven to eight after hatching at 16°C . Beating of the prototroch cilia can still be observed at this point, but the larvae lose their ability to swim before the trunk is covered by sclerites. This is different to our observation on W. argentea. Probably the relatively short prototroch cilia of the metamorphosing E. babai larvae render swimming more difficult. Epimenia babai juveniles lived up to one month after metamorphosis without food uptake , which despite of the difference in culturing temperature is similar to W. argentea.
The pericalymma larva of Solenogastres
The larva of Wirenia argentea is a pericalymma larva sensu Salvini-Plawen  or test cell larva sensu Heath  and Thompson , with a pronounced apical cap (preoral test). Okusu proposed homology of the somewhat reduced apical cap of Epimenia babai with the larger cap of other species  and we agree with this conclusion. In general, the posterior rim of the apical cap of solenogasters forms a ring-shaped fold (see Fig. 1J) and the relatively large cells lining the outer surface are commonly referred to as test cells. The general arrangement of the partly ciliated test cells stays the same until degeneration of the cap sets in. In W. argentea, the inner surface of the posterior rim of the apical cap - facing the trunk - is formed by smaller cells similar to epidermal cells of the trunk (Fig. 8A-F). No such information is available for any other solenogaster species. In some, such as Epimenia babai, the free part of the posterior apical cap rim is short and is reduced (stretched) at an early larval stage, thus exposing the blastopore [22, 23, 28]. In others, such as Wirenia argentea (described herein), Nematomenia banyulensis , and Neomenia carinata , the apical cap is more pronounced and the rim persists until the larva settles. Rhopalomenia aglaophaeniae  and Halomenia gravida  appear to present an intermediate stage, even if the status of the latter species is somewhat unsure because the brooded larvae detected in the mantle cavity of a sectioned adult do not represent the entire developmental sequence until metamorphic competence . The fact that H. gravida and R. aglaophaeniae, that have larvae bearing a well-developed apical cap, and E. babai with a reduced cap have been assigned to the supposedly derived order Cavibelonia [1, 31] points towards a tendency to cap size reduction within the clade.
Spiralian test cell larvae
The term "test cell larva" stems from early work on protobranch bivalve development and refers to the cover composed of large cells that entirely enclose the developing larval body [9, 10]. Test cells in general are cleavage arrested and deciduous, i.e., they are shed or absorbed during metamorphosis. Salvini-Plawen used the term pericalymma larva to describe a larval type characterized by a test-like cover (calymma) surrounding at least parts of the posterior larval body . He proposed homology of the larval tests between protobranch bivalve and solenogaster larvae, the so-called endolarva of polygordiid and oweniid polychaetes, and the sipunculan "serosa" larva. He furthermore suggested this lecitotrophic larval type to represent a more ancestral condition than the planktotrophic trochophore or veliger larvae [see also ]. There are, however, some conspicuous morphological differences between the diverse pericalymma larvae. The protobranch test is a preoral layer of large ectodermal cells enveloping the entire larval body. In older larvae it is separated from underlying tissues by a non-cellular space and thus only loosely connected to the remaining body [10, 32, 33]. Upon metamorphosis, this entire layer of test-cells is shed. Similar to the situation in Solenogastres, the apical-most cells are absorbed and in Nuculoida there is a ciliated postanal organ somewhat comparable to the posteriormost trunk area in Solenogastres . As an important difference to Solenogastres, the ecdodermal cell layer developing under the protobranch test forms the adult mantle epithelium and shell without any evertion process during metamorphosis. The protobranch test is entirely ciliated or bears three ciliary bands, but never has a single distinct locomotory band ("prototroch").
The overall morphology of larval test and ciliation is astonishingly similar in polychaete endolarvae and in solenogaster pericalymma larvae. In the first, however, the test cells are located preorally and postorally and also anterior and posterior to the prototrochal ciliary band, while in the latter all test cells are preoral and the distance between prototrochal cilia and oral opening is large. In general a metatroch appears to be lacking in mollusks (see below). The serosa or cellular embryonic envelope of the Sipunculus nudus (Sipuncula) larva appears to be a peculiarity of this single species [34, 35], and already Gerould proposed it to be secondarily derived from prototrochal cells . Based on morphological differences and considering the fact that according to recent phylogenies the serosa- and endolarvae do not occur in supposedly "basal" branches of Sipuncula and Polychaeta, it appears more probable that "test cell larvae" have evolved multiple times within Spiralia [see also [12, 23]]. Further knowledge about cell progeny is crucial to assess this hypothesis.
The finding of protonephridia in larval W. argentea is a novelty for solenogasters. These excretory organs are an important trait because their absence in marine mollusks has been used as one of the main arguments against the homology of annelid and mollusk trochophore larvae . Since then, larval protonephridia have been found in several mollusk taxa including caudofoveates  and chitons , and the evidence for larval protonephridia in W. argentea presented here confirms that these organs are ancestral for Mollusca and Trochozoa/Spiralia in general.
Trochs and other ciliation in Solenogastres
The apical cap of all hitherto described solenogaster larvae has a well-developed equatorial band of compound cilia, the prototroch sensu stricto [17, 22, 23, 28, 29]. Heath did not observe any ciliary bands in the brooded larvae of Halomenia gravida , but this might be due to the state of preservation of his material. In free-swimming solenogaster larvae, compound prototrochal cilia are generally restricted to a single row of cells. The cells of the remaining apical cap bear single cilia, which may be grouped into more or less diffuse bands, as for example in the post-trochal cell tiers of W. argentea.
Solenogaster larvae have a densely ciliated ring-shaped zone at the posterior end of their trunk, a telotroch, which is already present in very early larval stages and that prevails until metamorphosis [[23, 29], and herein]. It is not known if the compound cilia of the telotroch take part in locomotion. Interesting is the ciliation of the oral area and stomodaeum in W. argentea larvae, and the distinct troch surrounding the trunk apically (oral troch). Such ciliation has not been described before. Adult solenogasters have a preoral ciliary band surrounding the sensory vestibulum but in general lack locomotory cilia at the mouth opening and in the pharynx [38, 39]. The dense oral and stomodaeal ciliation in W. argentea larvae (especially in the earlier stages) could thus tentatively be interpreted as rudiments of a former planktotrophic feeding apparatus. The oral troch of W. argentea is located in a position comparable to the adoral ciliary zone or even to a metatroch in planktotrophic trochophore larvae (see below). In the recent, definitely non-feeding W. argentea larvae, the oral troch might serve to ventilate the ring-shaped cavity between the apical cap and the anterior trunk. The pericalymma larvae of Solenogastres in general are thought to be non-feeding, i.e. lecitotrophic , but in N. carinata Thompson shows the gut to be complete already at a very early stage [see figure 22 in reference ]. If these observations are correct, swimming stages or at least very young postlarvae of N. carinata may be capable of feeding. His text, however, diverges from the figures, because he describes the fusion between midgut and hindgut to take place during metamorphosis . This is at an early stage, compared to e.g., W. argentea or Epimenia babai [22, 28], but fits the general developmental patterns in lecitotrophic mollusk larvae .
The foot of W. argentea, similar to Nematomenia banyulensis  and E. babai , arises as a simple ciliary band ventrally on the trunk of the swimming larva and only later develops into a longitudinal ciliated fold that may be protruded or retracted by means of dorsoventral body muscles. Generally, the solenogaster foot - in contrast to the muscular foot of chitons and conchiferan mollusks - lacks intrinsic musculature and locomotion is largely due to ciliary action [ and own observations]. As stated for E. babai , there is no indication in development that the solenogaster foot represents a reduced muscular foot.
Homology of trochs in spiralian larvae
Cell lineage studies could confirm the homology of the prototroch across diverse spiralians, including chitons, polychaetes, and nemerteans [41–43]. Typically, prototrochal cells become cleavage arrested early in development, are thus larger than regularly dividing cells, and differentiate as ciliated cells . The definitions of prototroch cells and larval test cells thus are quite similar (see above). There is little doubt that the main locomotory troch in solenogaster larvae represents a prototrochal ciliary band comparable to that, for example, of gastropod veliger larvae or polychaete trochophores, but this prototroch s. str. of mollusks is merely locomotory in function . Information about cell progeny is necessary to infer if and to which degree the larval test of solenogasters and protobranch bivalves represents an enlarged prototrochal area. Cell lineage studies could also shed light on the potential homology of the oral troch of W. argentea to the adoral ciliary zone of other lophotrochozoan larvae or to a metatroch. It shall be noted here that the so-called "metatroch" of the gastropod Crepidula fornicata is probably not homologous to the metatroch of annelid trochophores [44, 45]. The ventral ciliation of the W. argentea larva - the future foot - resembles the early foot ciliation in gastropod and bivalve larvae, but also the locomotory neurotroch (gastrotroch) of certain polychaete trochophores [12, 45]. It is tempting to propose an evolutionary pathway leading from such larval ventral ciliation via a pedal fold, such as present in recent solenogasters, to the more complex muscular foot of "higher" mollusks.
Development of epidermal hard parts in mollusks
Our results on epidermal sclerite development demonstrate that in W. argentea dorsal sclerites do not show a serial arrangement in any ontogenetic stage. Pruvot's description of seven dorsal rows of imbricating scales in N. banyulensis  thus remains an isolated observation based on a single specimen. The situation described for an unidentified solenogaster postlarva  is different in that no single rows of sclerites were present, but an arrangement the authors termed "iterated dorsal groups of spicules". These six groups of sclerites appeared separated by bare areas that were proposed to be homologous to polyplacophoran shell fields. The entire sclerite arrangement of the postlarva was compared to the scleritome of Cambrian fossils, such as Wiwaxia corrugata and Halkieria evangelista. Together with this remarkable specimen, a second and slightly larger representative, supposedly of the same species, lacking the bare areas was found. The presence of very characteristic solid hook-shaped sclerites [see figure 1 in reference , photograph of the small postlarva] suggests an affinity to the probably derived solenogaster clade Phyllomeniidae . This phylogenetic status of the postlarvae somewhat weakens the evolutionary arguments presented , but the findings are of interest when considering the ontogeny of the scleritome (sclerite armour) in solenogasters. The merely 0.4 mm long postlarva had a stunning diversity of sclerite types and the sclerites were quite large in size compared to the total body length, as common in juvenile solenogasters .
It has been known for a long time [23, 30] that in representatives of the supposedly derived clade Cavibelonia the hollow acicular adult sclerites are preceeded by scale-like larval sclerites. The presence of scale-like sclerites has been interpreted as the ancestral status, thus rendering the order Pholidoskepia, characterized by a merely scaly scleritome, the most "basal" clade [5, 31]. Here, we show that also in Pholidoskepia species, such as W. argentea, the adult scales have larval precursors that are different in shape. This is an interesting analogous (?) case to shell development in gastropods with embryonic versus adult shells. In chiton larvae no such ontogenetic change in sclerite types is known and the mantle sclerites arise slightly earlier than the first seven shell plates, even though the shell plate anlagen are obvious as dorsal grooves already earlier than the sclerites [24, 47–50]. The sclerite and shell plate producing cells are of a more diverse and partly different cell progeny as compared to the shell gland cells of conchiferan mollusks: Conchiferan shells stem predominantly from the 2d and to a lesser extent from the 2a, 2b, 2c and 3c micromeres, while the spicule-producing cells of the chiton Mopalia mucosa stem from the 1a, 1d, 2a, 2d, and 3a micromeres and shell plates from the 2d, 3c, and 3d micromeres . In chiton larvae the most anterior part of the sclerite-bearing girdle appears anterior to the prototroch [41, 49] while in solenogaster larvae the cuticle- and sclerite-bearing epidermis is restricted to the area posterior to the main locomotory cilary band. Data on the progeny of solenogaster sclerite producing cells are necessary for further interpretation of this difference.