Comparison of present and previous studies of Oweniidae nervous systems
Recent phylogenomic studies of Annelida univocally show Oweniidae and Magelonidae (most likely together in the common clade Palaeoannelida) to branch off as sister group to the remaining groups of Annelida [11, 18].
Besides our studies of Myriowenia sp. and Owenia fusiformis we had the opportunity to look into unpublished sections of the nervous system in Owenia fusiformis and Magelona papillicornis done by Orrhage. We are aware that Orrhage’s Owenia species may not correspond to our O. fusiformis species . Despite this and the fact that his sections were stained with a different dye, they show the same details and results as our histological sections of O. fusiformis. This close correspondence justify a broader comparison of our results to the brain schemes set up by Orrhage on e.g. Magelonidae .
Phylogenetic relationships among Oweniidae have long been a matter of debate [41, 42]. The most recent study on oweniid in-group relationships  showed Owenia fusiformis and Myriowenia sp. to be nested in each of their different clades in the tree, making them good representatives for inferring the ancestral structure of the oweniid central nervous system and sense organs. In both species the central nervous system is basiepidermal, a position which is also known for other closely related Annelida like Magelonidae and Chaetopterifomia as well as for numerous pleistoannelid taxa [2, 11]. Thus, a basiepidermal central nervous system is proposed as representing not only the primary condition for Oweniidae but also as the ancestral state for annelids.
In both species studied the brain is a simple, basiepidermal ring that is continuous with the ventral nerve cord and shows no anterior split into a dorsal and ventral root of each lateral cord in adults. In Owenia fusiformis the ventral nerve cord arises from two perpendicular arranged lateral medullary cords, in contrast to Myriowenia sp. where the ventral nerve cord directly arises from the midventral part of the brain. Since Magelonidae also possess a pair of lateral medullary cords, arranged perpendicular to the brain , it is likely that two initially separated ventral medullary cords with non-segmental commissures, like in Owenia fusiformis, represent the primary condition. Therefore, a single ventral cord in Myriowenia must be a derived character state. Müller  stated that in the ground pattern of annelids the circumesophageal connectives have two roots. However, in presumably basally branching groups within annelids, such as Magelonidae, Apistobranchidae and Psammodrilidae circumesophageal connectives do not exist, since the whole cns is medullary (own unpublished observation). Moreover, the brain is ring-shaped, surrounds the terminal mouth and gives rise to the ventral cord without separate neurite bundles. This statement differs significantly from Orrhage’s extensive studies of the annelid nervous system in other groups ( for review). As mentioned above we also studied Orrhage’s original sections of Magelona papillicornis and were not able to reach the same conclusions as Orrhage , but this issue will be part of a follow- up investigation.
Regarding the presence of head appendages, all species of Myriowenia possess one pair of palps . Although unique among Oweniidae, they must be regarded as a synapomorphy of Palaeoannelida and the remaining annelids, since these paired feeding appendages which are found in most of the closely related annelid lineages, the so-called grooved palps, are assumed to be homologous across Annelida [1, 14]. However, a common origin of palps within Annelida has been debated since the first anatomical investigations on the annelid nervous systems were performed in the middle of the nineteenth century. Orrhage and Müller  give a comprehensive review on this topic (see also Gustafson p. 372 for historical background ). In both species studied here, two pairs of nerves branch off from the brain and enter either the palps of Myriowenia or enter into each half of the tentacle crown in Owenia fusiformis. These nerves do not only indicate homology of the tentacle crown of Owenia with the palps of Myriowenia, they also suggest two pairs of palp nerves in the oweniid stem species. Since magelonids and chaetopterids also possess a pair of palps with a similar innervation pattern , such types of palps are most likely the primary condition in Oweniidae and Palaeoannelida. Within Oweniidae they were either modified into a tentacle crown or were completely reduced in certain oweniid species such as Galathowenia and Myrochele. Therefore, in all probability this pattern must have been present in the last common ancestor of annelids as well and all other patterns observed represent derived character states.
A giant fibre is present in Myriowenia sp., but absent in Owenia fusiformis. Since magelonids and various other annelids  possess such fibres, their homology and secondary reduction in Owenia fusiformis seems to be the most parsimonious explanation.
Pigmented eyes are missing in Myriowenia sp., but are present in Myriochele and Owenia fusiformis and described for Galathowenia species [34, 43]. The structure of these eye spots differs from that present in other annelids e.g. [2, 12, 45]. Usually annelid eyes are cup-shaped and comprise two cell types: rhabdomeric photoreceptor cells and pigment cells with shading pigment. The phylogenetic significance of the eyes in Oweniidae is presently hard to evaluate. A detailed comparative study on the eye structure in oweniids and other annelid taxa is underway to clarify this question and to better understand the evolution of eyes in annelids.
Until now nuchal organs are regarded as one of the most important apomorphies of Annelida [14, 15]. Consequently, their absence in annelids has always been regarded as a loss and thus as secondary [12 for review]. However, in the light of the current molecular phylogenies this hypothesis must be reconsidered. Since our re-investigation verifies the absence of nuchal organs in adult Oweniidae and previous examinations support the lack of these structures in Magelonidae as well as in Chaetopteriformia , we hypothesize that this conspicuous sensory organ must have evolved within Annelida, in the stem lineage of the clade comprising Amphinomidae/Sipuncula and Pleistoannelida (Errantia + Sedentaria). Consequently, nuchal organs are not considered an apomorphic trait of Annelida (see also ).
Another type of prominent sensory organs – the lateral organs described for magelonids and other annelids are absent in Oweniidae. The laterally located, densely ciliated groove in Owenia fusiformis, which was described as a potential sensory structure by McIntosh , most likely has no sensory function since no connection of the ciliated cells to the nervous system was found in the present study.
Presently, there is no evidence that complex sensory organs such as elaborated eyes, nuchal organs or lateral organs were present in adults of the last common ancestor of annelids.
Simplicity in Oweniidae is not necessarily secondary
One could argue that the low complexity of the central nervous system in oweniid species and the lack of complex sensory organs results from the tube-dwelling mode of life and therefore, is secondary. If so, one would expect that the nervous system is also simple in those sedentary groups that are more derived within the annelid tree, like Sabellariidae or Sabellidae. Both comprise tube-dwelling species which should then accordingly have a comparable and structurally similar, “simple” nervous system. However, this is not the case. Although, the nervous system of e.g., Sabellaria alveolata shows less complexity compared to those of errant families such as Eunicidae, it still comprises a ganglionic, rope-ladder like ventral nervous system, a comparatively complex subepidermal brain and complex sensory structures such as nuchal organs [2, 11, 47, 48].
However, taking developmental studies into account the picture seems more complex. The brain of Owenia fusiformis initially shows several commissures in staining against anti- 5HT which fuse during development (, Fig. 4h). The same is true for Magelona filiformis (Helm, unpublished); however, these brain commissures are not retained as separate commissures in the adult stage. The annelid ancestor may likewise have had additional commissures during development, but there is no support yet for their presence in the ancestral adult stage. Moreover, it is unknown whether commissures in errant annelids like eunicids are incorporated from the larval stage into the adult brain or if these commissures develop de novo, emphasizing the need for more detailed studies into nervous system development to clarify these issues.
Given Orrhage’s reconstruction of the anterior nervous system in Magelona filiformis  was correct and we were simply not able to see his observation properly (unpublished study, not shown here), the presumed sister group of Oweniidae would have two roots of the circumesophageal connectives. In this case these structures may have been secondarily reduced in Oweniidae, except if Magelonidae turns out to be the sister group to the remaining Annelida (minus Oweniidae). This unresolved situation shows the necessity for further studies into the annelid nervous system on the one hand and a more thorough look into the potential outgroups on the other. In this respect a special focus should also be laid on the diversity of nerve cell types, as Magelonidae possess a higher diversity of nerve cell types than Oweniidae.
Comparison within Annelida and next to spiralian outgroups
For discussion of the evolution of the nervous system of Annelids we use the phylogeny presented by Helm et al. . Since it still is a matter of debate which lophotrochozoan taxon actually constitutes the sister group of annelids, we discuss characters of several of the putative next relatives among Spiralia, such as Nemertea, Mollusca, Phoronida, Bryozoa, and Brachiopoda [19, 49, 50].
Position of the cns. A basiepidermal position of the nervous system such as present in oweniid species is also found in certain but not all taxa of Sedentaria (e.g. Cirratulidae) and Errantia (e.g. Nephtys hombergii) [2, 3]. A subepidermal position of the nervous system is found in taxa of the Eunicida (e.g. Marphysa bellii) and in some taxa of the Sedentaria (e.g. Sabellaria alveolata) [3, 11]. In species representing sister taxa to remaining taxa within related lophotrochozoan lineages such aslike Nemertea , Brachiopoda and Phoronida the central nervous system is also basiepidermal [3, 52, 53]. In Mollusca the nervous system is exclusively subepidermal (intramuscular) [3, 54–57]. Using the most recent phylogenies of Spiralia as backbone [49, 50], the basiepidermal position of the cns most likely represents the plesiomorphic condition in Spiralia and indicates that the shift of the nervous system beneath the basal lamina of the epidermis into the mesodermal tissue (musculature) occurred secondarily and repeatedly within the different spiralian taxa.
Brain. In the Oweniidae species investigated the brain is circular and surrounds the mouth opening. There are no dorsal enlargements, lobes or commissures/tracts. Neither a supraesophageal nor a subesophageal ganglion (or any dorsal clusters of somata) is present; ganglia were also not found in the ventral nerve cord. The homogenous distribution of somata forming a crescent layer around the neuropil classifies the entire central nervous system as medullary and not ganglionic . Posterior to the mouth opening the neuropil of both body sides fuse to form a single cord. This morphology is the simplest with respect to somata distribution and the enlargement of the dorsal region of the brain found in Annelida (see also [2, 34, 38]).
In other spiralian taxa like phoronids the cns is simple, consisting of a dorsally located brain with no enlargements and an epidermal nerve plexus in the remaining body. Ganglia and polymorphic neuron clusters are not reported . In Brachiopoda the brain is also simple with respect to somata distribution and dorsal enlargements . In Broyozoan the cns is a neuroepithelium and polymorphic neurons are reported. [58, 59]. The most basally branching taxon of Nemerteans also possess a circular brain, a medullary cns and only slightly dorsal enlargements of the brain [51, 60]. In mollusc taxa Caudofoveata, Solenogastres and Polyplacophora the cns is also medullary, the brain is more or less circular with no dorsal enlargements and gives rise to two lateral (visceral) and two ventral (pedal) medullary cords [54, 61]. All these findings indicate that a simple medullary, ring-shaped brain represents the ancestral condition in annelids, and most likely also a plesiomorphy of Annelida as well as several other spiralian lineages.
Within Annelida the dorsal aspect of the brain is enlarged by evolving ganglia, additional fiber cores in the brain and clusters of polymorphic neurons. This also applies for other taxa, like Nemertea and Mollusca [3, 60, 62]. Along with such an expansion of the brain, higher brain centers such as the mushroom bodies, evolved convergently within Annelida and presumably other metazoan groups . Evolution of brain complexity within different spiralian lineages might explain traditional difficulties to homologize certain brain areas across Spiralia . Based on the recent phylogenies higher brain centers most likely evolved independently in different taxa, so that any attempt homologizing them is futile.
In Oweniidae only one type of neurons prevails in the central nervous system. In contrast to other neurons this small type of neuron possesses a very small cell body around the nucleus. . Additional neuron types evolved in the stem lineage of Pleistoannelida+ Sipuncula/Amphinomidae [2, own unpublished observation]. In basally branching lineages within Nemertea  and Polyplacophora ( for review,) only one class of neurons is also present, while the number of different polymorphic neuron classes increases independently within all larger spiralian groups, including Annelida [3, 60, 62].
Glia. Conspicuous are the intermediate filaments which run through the neuropil of the cns of both species investigated. These filaments form bundles in glial cells and are attached to the basal lamina of the epidermis via hemidesmosoms. These glial cells are called fibrous glia  or radial glial cells . According to their position and ultrastructural details radial glial cells represent modified epidermal cells. In contrast to the latter, they are able to secrete a protein called SCO-spondin which contributes to the formation of Reissner’s fibre . Glia cells are distinguishable from epidermal cells, since they possess gliosomes and dense bundles of intermediate filaments. In epidermal cells intermediate filaments are also present and attach these cells to the basal lamina, but do not form such prominent bundles. Intermediate filament bundles are thick, twisted yellow bundles in Azan stained histological sections. While they are clearly visible in the nervous system they are not present in the epidermis. The same kind of cells are also found in Phoronida , Nemertea  and Polyplacophora (own unpubl. observation) indicating an early evolution of these cells. Moreover, in errant annelids glial cells form a cortex around the neuropil layer [2, 68, 69]. Baskin regarded this massive layer of glial cells to be important to prevent mechanical stress to the nervous system while moving [68, 69]. In so far investigated species representing sister clades to remaining clades within the major Spiralian taxa, a glial- cell layer surrounding the brain is not present. Radial glial cells with a prominent intermediate filament system seem to be especially important in species which possess a basiepidermal nervous system, since the epidermal cells which overlie the neuropil have to be attached to the basal lamina. Additionally, this basiepidermal fiber system is important to maintain the structure of the epidermis and the underlying cns.
As outlined above, the so-called eye spots in oweniids are presently difficult to evaluate due to their absence in some oweniid species and missing information on the eye structure in other basally branching annelid taxa. The eyes of Owenia fusiformis do not correspond currently to any type of eye found in the remaining annelids and are missing in the other oweniid taxon investigated, Myriowenia. Structurally these eyes somewhat resemble those of larval polyplacophorans or some gastropods such as limpets in consisting of a simple monolayer of epidermal pigment- and sensory cells .