The “Spaghetti Project”: the final identification guide to European Terebellidae (sensu lato) (Annelida, Terebelliformia)

. This paper is the conclusion of the “Spaghetti Project” aiming to revise French species of Terebellidae sensu lato (s.l.) belonging to the ﬁ ve families: Polycirridae, Telothelepodidae, Terebellidae sensu stricto (s.s.), Thelepodidae and Trichobranchidae. During this project, 41 species were observed, 31 of them new for science: eight species of Polycirridae, eleven species of Terebellidae s.s., three species of Thelepodidae and nine species of Trichobranchidae. We provide a comprehensive key for all European species of terebellids with a focus on the important diagnostic characters for each family. Finally, we discuss issues on taxonomy, biodiversity and cryptic and pseudo-cryptic species of polychaetes in European waters, based on results obtained during this project. 2021. The “Spaghetti Project”: the ﬁ nal identi ﬁ cation guide to European Terebellidae (sensu lato) (Annelida, Terebelliformia). European Journal of Taxonomy 782: 108–156.

These tubiculous polychaetes are characterised by the presence of numerous grooved buccal tentacles used for selective deposit feeding, rendering these animals the name of "Spaghetti worms" (Hutchings et al. 2021b). These tentacles are of prostomial origin and not retractable into the mouth. They are generally smooth, except for in some polycirrids where they are papillose. Most of the terebellids are sedentary worms found in all marine environments, from the intertidal to the abyss and are common worldwide, distributed from polar to tropical regions (Hutchings et al. 2021b). The fi ve families belonging to Terebellidae (s.l.) can be separated from each other by the morphology of the upper lip, the shape and number of branchiae, the glandular areas of ventral segments, the neuropodia and the arrangement of the uncini of anterior segments (i.e., in single or double rows) (Hutchings et al. 2021b).
The "Spaghetti Project" was initiated when the fi rst author realized that the taxonomy of these worms in France, but also in Europe, was poorly documented. Indeed, the lack of accurate literature and the absence of useful and up-to-date keys of identifi cation for this part of the world has led to their misidentifi cations for decades. In 2016, after observations of several specimens of Terebelliformia during a national workshop (at Arcachon, conducted by Mario Londoño-Mesa) and a national taxonomic course (at Caen, conducted by Pat Hutchings), we realized that many of the species required in-depth investigations. This collaborative project involved all benthic taxonomists at all French marine stations (RESOMAR network) who sent us fresh material as well as specimens stored in local collections. The fi rst part of the project, devoted to the Trichobranchidae, allowed us to describe nine new species along the French coasts (Lavesque et al. 2019a). The second paper, focused on Telothelepodidae and Thelepodidae, described three new species (Lavesque et al. 2020a) and the third one on Polycirridae described eight new species (Lavesque et al. 2020b). Finally, the fourth paper dealt with the Terebellidae sensu stricto (s.s.) and included the description of nine species (Lavesque et al. 2021). With the previous descriptions frequently extending ventrally, terminating laterally to mouth ( Fig. 2A-D). Buccal tentacles of two types at least, short ones thin, uniformly cylindrical, long tentacles stouter, expanded at tips to variable degrees, distally spatulate (Fig. 2B, D) or more specialised. Peristomium forming lips; lips expanded, upper lip large, frequently circular and convoluted, folded into three lobes; swollen lower lip, only midventral or cushion-like across ventrum, sometimes extending posteriorly for a few segments ( Fig. 2A-D). Segment I reduced, frequently only visible ventrally, sometimes completely hidden. Segment II distinctly narrower than following segments, constricting body posteriorly to "lips head"; SG II usually with rectangular or pentagonal mid-ventral shield at beginning of mid-ventral groove, sometimes extending anteriorly through SG I until near posterior margin of lower lip (Fig. 2C). Anterior segments highly glandular ventrally, frequently papillose or tessellated, with paired ventro-lateral pads separated from each other within pairs by mid-ventral groove extending from SG II-IV to posterior body ( Fig. 2A-D). Branchiae absent. Notopodia, if present, from SG III ( Fig. 2A-D), extending for variable number of segments, usually few; bilobed, elongate notopodia, post-chaetal lobes sometimes longer, notochaetae originating between lobes along all extension of notopodia, separating lobes from base on ventral side of notopodia ( Fig. 2A-D); notochaetae winged ( Fig. 2E) and/or pinnate, wings of variable width. Neuropodia, if present, located posteriorly to notopodia, frequently from posterior thoracic segments or only on abdomen; neurochaetae as acicular spines or avicular uncini, of two types, and arranged in a single row (Figs 1C,. Nephridial and genital papillae usually present, at anterior bases of all notopodia, or only at anteriormost notopodia ( Fig. 2A). Pygidium smooth or with rounded ventral papilla.

Main morphological characters of European species
PARAPODIA. The parapodia of the members of this family are extremely important to separate the different genera. The genus Hauchiella is characterised by the absence of parapodia and Lysilla by the absence of neuropodia only. The neuropodia of members of Amaeana are characterised by the presence of spines, while those of Polycirrus bear avicular uncini (Figs 1B,. Within the genus Polycirrus, the number and location of segments with notopodia and/or neuropodia are of important taxonomic value. Particularly, some species have uncini present only on abdominal segments, i.e., on segments without notopodia, and others have uncini starting before the end of the thorax, on segments bearing also notopodia. SHAPE OF THE LIPS. As for other terebellids, polycirrids have a peristomium with well-defi ned upper and lower lips. The upper lip is large and can be trilobed (Fig. 2B) or with a single medial lobe (Fig. 2D). Generally, the upper lip is trilobed but the lobes differ in size and shape and lateral lobes can be reduced or well developed. The shape and the size of the lower lip is also highly variable between species. This lip can be rectangular, squared, rounded or subtriangular, swollen or not, longer than wide or wider than long ( Fig. 2B-D).   (Malm, 1874) Sweden GNM NOTOCHAETAE. Two types of notochaetae can be present: winged chaetae as for P. glasbyi (Fig. 2E) and/ or pinnate as for P. plumosus. The winged notochaetae have wings of different width which are often conspicuous under light microscope but appear hirsute under SEM (Fig. 2E).
UNCINI SHAPE AND DENTICULATION. In Polycirrus two types of uncini are present: Type 1 with a short occipitum (back) and a straight to slightly convex base (Fig. 1B); and Type 2 with a long occipitum and a concave base (Glasby & Hutchings 2014). To date, all described European species have Type 1 uncini. The denticulation of uncini is also helpful in separating species, with the presence (as for P. catalanensis) (Fig. 2F) or the absence (as for P. arenivorus) of a main tooth above the main fang, and the number of rows of secondary teeth.  Transverse prostomium attached to dorsal surface of upper lip; basal part as thick crest, eyespots frequently present in one pair of dorso-lateral clusters, each with several rows of eyespots (Fig. 3A); distal part at base of upper lip, frequently with low or erect mid-dorsal tongue-like process, fused to upper lip at variable degrees, with free distal lobe(s), or free from the base. Buccal tentacles of two types, short ones thin, uniformly cylindrical, long tentacles stouter and expanded at tips, slightly spatulate (Figs 3A-B, F, 4A). Peristomium forming lips and continuing dorsally at least for short extension, with dorso-lateral nuchal organs at margin with prostomium; lips expanded, upper lip distinctly elongate and narrow, undulated to convoluted; swollen lower lip extending across ventrum, cushion-like or segment-like, frequently deeply grooved (Figs 3A-B, 4A). Either SG I or SG II reduced, not forming complete ring in many species. Anterior segments glandular ventrally, smooth, discrete shields absent and frequently with glandular regions poorly developed in comparison to other families of Terebellidae s.l.; mid-ventral groove frequently extending from anterior segments. Two pairs of cirriform branchiae on SG II-III, each pair with simple thin, curled and relatively short fi laments progressively tapering to tips (Figs 3A, 4A), originating from raised crests on anterior margins of SG II and III, or from specialised, apparently glandular, dorso-lateral cushion-like pads occupying from anterior margins to level of posterior bases of notopodia of those segments. Notopodia beginning from SG II or III, usually SG III, extending for at least 15 segments; notopodia as short cones, notochaetae originating from central core on top, distal lobes absent; notochaetae winged, sometimes with bulbous head and alimbate tips (bayonet-like chaetae), at least in anterior row of anterior thoracic segments. Neuropodia beginning posteriorly to notopodia, usually around SG VIII-XII; neuropodia in conjunction with notopodia as sessile tori, as distinctly low pinnules after notopodia terminate; neurochaetae in single row, as avicular uncini about as long as high, with short triangular heel directed posteriorly, wide and slightly curved base, and dorsal button near midlength of uncini, but closer to anterior margin (Fig. 4E). Nephridial and genital papillae, if conspicuous, on SG V-VII, posterior to bases of notopodia.

Remarks
This recent family was described by Nogueira et al. (2013) after conducting a comprehensive phylogenetic analysis. The members of this family were previously considered as Thelepodidae but differ in having a narrow and elongate upper lip, poorly developed neuropodia and anterior segments less glandular ventrally than in other thelepodids. In European waters, this family is represented by a single species, Parathelepus collaris ( Transverse prostomium attached to dorsal surface of upper lip; basal part as thick crest, eyespots frequently present, in short lateral rows, or extending transversely across basal part of prostomium, usually progressively more spaced towards dorsal mid-line, with mid-dorsal gap or not; distal part of base of upper lip short, from nearly indistinct to shelf-like. Buccal tentacles all uniformly thin and cylindrical, to slightly spatulate distally (Figs 3D, F, 4B). Peristomium forming lips, sometimes also complete annulation, with dorso-lateral nuchal organs as ciliated grooves; lips expanded, relatively short upper lip, hood-like, about as long as wide; swollen, button-like, mid-ventral lower lip (Figs 3D, F, 4B-C). Segment 1 usually present all around, frequently with ventral lobe marginal to mouth (Figs 3D, F, 4B-C); SG II typically with anterior margin as protruding crest, at least ventrally (Figs 3D-E, 4B-C); lobes on following anterior segments sometimes present. Anterior segments highly glandular ventrally, smooth to highly corrugated between neuropodia within pairs, discrete shields absent (Figs 3D F, 4B); mid-ventral groove frequently extending from anterior segments with notopodia. Two to three pairs of branchiae, on SG II-III or II-IV, each pair with simple thin, curled and relatively short fi laments progressively tapering to tips (Figs 3C, E, 4C), leaving mid-dorsal gap or not between fi laments within pairs; branchial fi laments originating directly from the body wall or from specialised dorsolateral cushion-like pads. Notopodia beginning on SG II-III, usually extending to mid-body, at least, sometimes until near posterior end; cylindrical to rectangular, distally bilobed notopodia, notochaetae originating between lobes; most taxa with winged notochaetae only, with wings of variable width ( Fig. 4D), distally serrated notochaetae sometimes also present; bayonet-like and pinnate chaetae both absent. Neuropodia beginning posteriorly to notopodia, on SG IV-VI, typically on SG V; neuropodia in conjunction with notopodia as fl eshy, swollen ridges, as raised rectangular to cylindrical pinnules after notopodia terminate; neurochaetae as avicular uncini frequently longer than high, with short triangular heel directed posteriorly, distinctly curved and wide base, and dorsal button near anterior margin of uncini, or within anterior third of distance between anterior margin of uncini and base of main fang ( Fig. 4F). Nephridial and genital papillae usually present, on SG IV-VII, posterior to bases of notopodia or between parapodial lobes (Fig. 3C).

Main morphological characters of European species
BRANCHIAE. Both in Thelepus and Streblosoma genera, the number of pairs of branchiae varies between two (e.g., Streblosoma lindsayae or Thelepus nucleolata) and three (e.g., Streblosoma hutchingsae or Thelepus setosus). Branchiae in Thelepodidae are always cirriform (Figs 3C, E, 4C) but the number of branchial fi laments varies among the species with for example 5-10 fi laments on the second and third pairs of branchiae for Streblosoma cabiochi (Fig. 3E) and only three or less fi laments for Streblosoma intestinale. Finally, the size of the medial dorsal gap separating the pairs of branchiae is a good diagnostic character. This gap is for example inconspicuous for T. parapari and wide for Thelepus cincinnatus (Nogueira 2019).

PRESENCE OF EYESPOTS. The eyespots are very useful in differentiating species of Streblosoma and
Thelepus for which they can be absent (e.g., Thelepus davehalli or Streblosoma hutchingsae) or present (e.g.m Thelepus corsicanus or Streblosoma nogueirai). Also, the arrangement of the eyespots, if in a continuous line, or leaving a medial gap is of taxonomic importance (Nogueira et al. 2010).
START AND EXTENSION OF NOTOPODIA. The segment with the fi rst appearance of notopodia permits the discrimination between the genus Streblosoma, for which notopodia begin on the second segment, and Euthelepus and Thelepus for which it begins on the third segment. These notopodia also extend for a variable number of segments, sometimes present only on the anterior half of the body (e.g., T. corsicanus) or present until the end of the body (T. japonicus). In most of the species, the uncini start on SGV which could correspond to CH3 (as in Thelepus) or CH4 (as in Streblosoma). The uncini are arranged habitually in single rows but some have uncini forming loops (C-shaped arrangement) from mid thorax onwards. This last character is found for example in S. nogueirai. Between species, the uncini differ in the development of the prow (e.g., well developed in T. triserialis), the shape of the base (e.g., strongly curved in S. cabiochi), the position of the dorsal button (e.g., far from anterior margin in S. bairdi or in a terminal position for T. japonicus (Fig. 1F) and number of secondary of teeth.
CREST  Transverse prostomium attached to dorsal surface of upper lip; basal part as thick crest, eyespots frequently present (Fig. 5B), in short rows at each lateral sides, or extending transversely across basal part of prostomium. Buccal tentacles all usually uniformly cylindrical. Peristomium usually forming lips only; lips expanded, relatively short upper lip, hood-like, about as long as wide; swollen, usually button-like and mid-ventral lower lip. Segment I terminating laterally to ventro-laterally, partially fused to expanded lower lip, or developed, forming lobes of variable extension and position. Lobes on anterior segments frequently present, of variable length, sometimes extending to SGV-VII (Figs 5B-D, 6A-D). Anterior segments highly glandular ventrally, with discrete, smooth to corrugated, rectangular to trapezoidal mid-ventral shields extending from anterior segments until termination of notopodia, or near it; mid-ventral groove extending from termination of mid-ventral shields. Two to three pairs of branchiae usually present (Figs 5A-D, 6A-D), but three genera have a single pair and several are abranchiate; branchial fi laments originating all together from single point on body wall, on either side of branchiferous segments, unbranched, or, more frequently, originating from conspicuous main stalk on either side of pair, branching from one to several levels, in plumose (spiraled), dichotomous, pectinate or arborescent arrangement. Notopodia beginning on SGII-V, SGIV in most genera, usually extending to mid-body, around SGXX, but sometimes present on fewer segments or extending more posteriorly for variable extension, rarely until near posterior end; fi rst pairs of notopodia inserted dorso-laterally, progressively more laterally, then vertically aligned; cylindrical to rectangular notopodia. Notochaetae originating from central core on top, distal lobes absent; notochaetae distally winged, wings of variable length and width, or serrate, sometimes with wings at midlength, basally to a serrated blade; some more specialised types of notochaetae may be present (Fig. 5E-G). Neuropodia beginning posteriorly to notopodia, on SGV-IX, usually on SGV; neuropodia in conjunction with notopodia as low, sessile ridges, sometimes continuing posteriorly until pygidium, but most taxa with rectangular to cylindrical or foliaceous neuropodial pinnules after notopodia terminate; neurochaetae as avicular uncini usually as long as high, with short triangular heel directed posteriorly, slightly curved and wide base, and dorsal button (Figs 1C-D, 6E-F); uncini arranged in double rows (Fig. 6E) from around SGXI usually until termination of notopodia, but several genera with double rows.

Remarks
In European waters, the Terebellidae s.s. are represented by 19 genera and 44 species (Table 1) (Fig. 5D). Generally branchiae are branching (dichotomous or arborescent), originating dorsolaterally from a main stalk ( NEPHRIDIAL AND GENITAL PAPILLAE. Terebellids are characterised by the presence of papillae situated close to the notopodia or between parapodial lobes. The nephridial papillae occur from SG III-V, while genital papillae are present from SG VI onwards and are prominent only when the animals are mature ( Fig. 5C-D, 6C). When they are visible, the morphology and the number of these papillae and their number permit the discrimination of species, as for Amphitrite or Terebella for example.
NOTOPODIA AND NEUROPODIA. Terebellidae s.s. differ by the number of pairs of notopodia, the segment on which notopodia and neuropodia start and the morphology of both noto-and neurochaetae. Usually, notochaetae are present on 17 segments, beginning from SG IV, but several exceptions exist as for example for Lanassa (n < 15) or Terebella (n > 25, often present to the pygidium). Notochaetae of Terebellidae are divided in two types: distally smooth as in Pista, Eupolymnia or Lanice, or distally serrated as in Amphitritides or Paramphitrite (Fig. 5E), and each types are sub-divided in sub groups (Nogueira et al. 2010: table 4). Concerning the neurochaetae, each part of the uncinus (Fig. 1C-D) differ greatly among the genera of Terebellidae and their morphology should be examined in detail. For example, members of the genus Pista have long-handled uncini, with the handle originating from the heel (Fig. 1C) vertically in a single row (Fig. 1D), while other species have multiple transverse rows of secondary teeth above the main fang ( Fig. 6E-G). The dorsal button is generally well developed and situated about midway between the base of the main fang and the tip of the prow, but is inconspicuous in specimens of Lanice and can be closer to the tip of the prow, as in Eupolymnia gili or the base of the main fang, as for Artacama proboscidea. Finally, the prow and the heel vary in shape and can be distally rounded or pointed.    (Risso, 1826) (Grube, 1860) 42  Transverse prostomium attached to dorsal surface of upper lip; basal part as thick crest, eyespots sometimes present; distal part at base of upper lip or extending along lip. Buccal tentacles of two types, uniformly cylindrical and expanded at tips, spatulate. Peristomium forming lips, sometimes also a ventral lobe, as an extension of the lower lip; lips expanded, circular upper lip, distal margin folded or convoluted; lower lip button-like, usually continuing by ventral lobe, or expanded, forming large scoop-shaped process (Figs 7A-C, 8A, C-D). Segment I usually short, frequently only visible ventrally; anterior margin of anterior segments with lobes as low, even-length collars covering posterior margins of preceding segments, at least ventrally; ventro-lateral or lateral lobes on anterior segments sometimes present. Anterior segments poorly glandular ventrally, smooth, discrete shields absent; midventral groove extending from posterior segments with notopodia. Two to four pairs of branchiae, beginning from SGII, each pair with single, thick and elongate, tapered or foliaceous fi lament, or two pairs fused in single four lobed structure originating mid-dorsally between SGII-III or II-IV (Figs 7C, 8C-D). Notopodia beginning from SGIII-VI, typically terminating at SGXX; short, conical notopodia, chaetae emerging from central core on top, distal lobes absent; narrowly-winged notochaetae in both rows throughout. Neuropodia beginning on same segment as notopodia or slightly posteriorly, rarely beginning before notopodia; sessile neuropodia until termination of notopodia, neurochaetae emerging directly from body wall, as rectangular to foliaceous pinnules after termination of notopodia; thoracic neurochaetae as acicular uncini (Figs 1A, 7D, 8F), sometimes with small hood or beard below main fang; avicular abdominal uncini, with secondary teeth in rows on top and laterally to main fang. Nephridial papillae on SGIII usually present, other papillae sometimes present on SGVI and SGVII, but reduced to inconspicuous in most taxa. Pygidium smooth to slightly crenulate, sometimes bilobed.

Remarks
In the past, the Trichobranchidae family was considered to be a subfamily of Terebellidae (Fauvel 1927;Day 1967;Garrafoni & Lana 2004), but recent phylogenetic analyses support the hypothesis of a valid family (Glasby et al. 2004;Nogueira et al. 2013). The family includes only three genera, i.e., Octobranchus Marion & Bobretzky, 1875, Terebellides Sars, 1835, and Trichobranchus Malmgren, 1866. For Trichobranchus and Octobranchus, only three species of each occur in Europe. The genus Terebellides is very speciose and is represented in Europe by 19 species, 13 of them described in the last two years (Lavesque et al. 2019b;Parapar et al. 2020a) (Table 1).

Main morphological characters for European species
The number of branchiae is the best character to discriminate the different genera, with Terebellides having a single large branchia, Trichobranchus with two or three pairs of branchiae and fi nally Octobranchus with four pairs.
Trichobranchus species are easy to differentiate based on the number of branchiae (two vs three) (Figs 7C, 8C) and the absence or presence of eyespots. In Octobranchus, the species differ by the shape of the branchiae (Fig. 8D) and the number of secondary teeth above the main fang of the uncini. Regarding Terebellides species, recent studies highlighted that several characters are very important for identifi cation to the species level (Lavesque et al. 2019a;Parapar et al. 2020aParapar et al. , 2020b. However, as many cryptic species occur at a small geographical scale (Nygren et al. 2018), which currently are confi rmed only by molecular analyses (Parapar et al. 2020a) much more work needs to be done to resolve all the species present.
BRANCHIAE. Even if Terebellides branchiae seem to be very similar within the genus (Figs 7A-B, 8A-B), several morphological characters permit the discrimination of species, such as the presence of a fi fth anterior branchial lobe (e.g., T. europaea), the degree of fusion of both upper and lower lobes (e.g.. not fused on T. ceneresi), the presence of long terminal fi laments (e.g., in T. shetlandica) or short posterior processes (Fig. 7B), and fi nally the presence and the shape of papillae situated on the margins of the branchial lamellae (Fig. 8B) (e.g., T. lilasae).
NOTOCHAETAE FROM FIRST CHAETIGER. The size of notochaetae of the fi rst chaetiger varies between species. For most of the species, these chaetae are of a similar size compared to those of the following chaetigers. However, they can be absent or much shorter (e.g., T. ceneresi) or much longer (e.g., T. mediterranea).
PRESENCE OF GENICULATE CHAETAE ON ONE OR TWO CHAETIGERS. The geniculate chaetae are exclusive to members of Terebellides and they are typically present on CH6 (SG VIII) only (Fig. 8E), but in some species they are present on two chaetigers, as for example in T. bigeniculatus.
UNCINI DENTICULATION. The different types of uncini follow the classifi cations provided by Parapar et al. (2020b) for thoracic uncini (Fig. 8F) and Parapar et al. (2020a) for abdominal uncini. These classifi cations are based on the ratio between the length of the main fang (rostrum) and the crest of secondary teeth (capitium), and the size and number of the secondary teeth.
THORACIC CILIATED PAPILLAE. Following the recent study of Parapar et al. (2020a), the absence or the presence of thoracic ciliated papillae allow for the discrimination of Terebellides species. These papillae are situated dorsally to the thoracic notopodia (see for example Parapar et al. 2020a; Fig. 7B).
METHYL GREEN PATTERN. The colouration of Terebellides specimens prior to identifi cation is essential. Indeed, MG staining highlights the presence and the shape of the glandular region of the third thoracic chaetiger (e.g., undulating glandular region present and in members of T. gentili, oval for T. lilasae

Why have so many new species been discovered in such well-known waters?
In Europe (Greenland included), 109 valid species had been described since the fi rst description of Lanice conchilega by Pallas (1766). Most of these species (i.e., 44) were described by early European taxonomists in the 18 th and 19 th centuries, and only a few during the 20 th century (12 species). However, only fi ve species were described before the start of this project from French waters: Amphitrite edwardsii, Pista mediterranea, Polycirrus arenivorus, Polycirrus denticulatus and Thelepus setosus. In addition, four species described from French waters are now considered as nomen dubium: Lanassa proecox (Saint-Joseph, 1899) which could be a postlarval stage of a known species (Fauvel 1927;Gil 2011), Polycirrus haematodes (Claparède, 1864) and Polycirrus pallidus (Claparède, 1864) for which no type material exists and the original descriptions are very brief (Glasby & Hutchings 2014), and fi nally Amphitrite ramosa Risso, 1826, stated to be indeterminable based on the original description (Jirkov 2020). Since the start of the "Spaghetti Project" in 2018, more than 400 specimens were carefully examined and more than 100 molecular sequences obtained. In French coastal waters, 58 species occur, 31 of them described as new during this project. The fi rst question we can ask ourselves is: why? Why have so many new species been discovered in such well-known waters?
How can we explain the quasi-absence of discovery of new species in France for over a century? The fi rst reason is the diffi culty to identify known European terebellids. Indeed, as commented on by Hutchings & Lavesque (2020), the lack of literature and type material are especially challenging for taxonomists. Most of the European species were described by earlier workers who failed to designate type specimens and to deposit them in an offi cial collection, or when they did, material is often damaged and unusable (Lavesque et al. 2021). Moreover, they provided only approximate type localities and few details on habitat preferences. Thus, comparison between new material and type material is diffi cult. Referring to original descriptions is not helpful either; they are usually very brief with inadequate fi gures, and could correspond to several species because of the lack taxonomic details (Hutchings et al. 2021a(Hutchings et al. , 2021b.
The second reason, without any mystical connotation, is linked to the spectre of the priest Pierre Fauvel. Actually, the main reference work in polychaete taxonomic literature is, without any doubt, his "Faune de France" (Fauvel 1923(Fauvel , 1927. These two books are widely used by taxonomists, ecologists, students and private companies in France but also worldwide (Hutchings & Kupriyanova 2018;Hutchings & Lavesque 2020;Capa & Hutchings 2021). Fauvel was one of the most prolifi c authors in the history of polychaete taxonomy with 141 accepted species described, ranking 16 th polychaetologist in the world (Pamungkas et al. 2019). Surprisingly, he described only four species from French coastal waters as most of his works were focused on the fauna from India (e.g., Fauvel 1932) or Africa (e.g., Fauvel 1918). He also described many deep sea species which are stored in the Musée Océanographique de Monaco and were sampled in the European Atlantic Ocean during the "Hirondelle" (1885-1888) and the "Princesse-Alice" (1894-1897) cruises by the Prince Albert 1 st (e.g., Fauvel 1913). The specimens examined by Fauvel for his "Faune de France" were collected mainly at low tides or during dredging campaigns, while only some were received from a few colleagues (Fauvel 1923). In comparison, we had the opportunity to examine specimens from a greater variety of habitats, thanks to our RESOMAR colleagues working in eight coastal laboratories along the French coasts. With their help, we were able to compare material from a wide range of habitats, depths, and ecosystems. For example, 12 species were described from maerl (rhodolith) beds in Brittany (Lavesque et al. 2019a(Lavesque et al. , 2020a(Lavesque et al. , 2020b(Lavesque et al. , 2021, confi rming that this habitat is an important hotspot of biodiversity Barbera et al. 2003). Moreover, our colleagues also undertook new sampling excursions to obtain fresh material so that we could undertake molecular analyses. These analyses, coupled with morphological observations, permitted us to confi rm the existence of many cryptic species belonging to several species complexes such as the "Terebellides stroemii complex", "Pista cristata complex" or "Eupolymnia nebulosa complex" (Lavesque et al. 2019a(Lavesque et al. , 2021. Even if, as taxonomists, we work in a similar way to Grube, Malmgren, McIntosh and other early scientists spending hours behind a stereo microscope, we are fortunate to have access to advanced technologies like high resolution cameras, scanning electron microscopes, molecular laboratories and internet facilities. These technologies help us to fi nd differences or characters that early taxonomists would have missed and easy access to all the available literature. The third reason is the lack of accurate literature for European waters, which is intimately linked to Fauvel's work. His two volumes of the "Faune de France" were of an excellent standard for his time. But publication was time consuming and costly, and resources were lacking to update his work in subsequent decades. For a long time, to 'correctly' identify a terebellid worm from French waters meant using either Fauvel's or Holthe's books. The latter, more recent work (Holthe 1986) is based on accurate observations (type material when possible), but the diagnoses are very short and do not take into account recent valuable taxonomic characters. Moreover, this work was focused on Scandinavian waters, from Greenland to Great Britain, a large region which differs from French waters and other countries from southern Europe especially with regard to water temperatures. Fauvel's books were widely used for nearly a century in France, in Europe and also in the rest of the world. This wide use was not a major issue for decades as scientists, polychaete taxonomists in particular, were convinced by the cosmopolitanism of marine worms (Hutchings & Kupriyanova 2018). Kristian Fauchald was the fi rst one to suggest that polychaetes can show interesting biogeographical patterns when properly identifi ed (Fauchald 1984). Recent studies clearly confi rmed that species of polychaetes have restricted distributions and this is particularly true for terebellids. Focusing on the genus Terebellides in Northern European waters, Nygren et al. (2018) identifi ed more than 25 species hidden within the so-called "cosmopolitan" species Terebellides stroemii. Most of these species occur only in a restricted area and specifi c habitat, and two species from Northern Europe are confi rmed for French waters: T. europaea and T. scotica (Lavesque et al. 2019a;Parapar et al. 2020a).
The fi nal explanation comes from the lack of taxonomic positions in France. This country is known for its famous early taxonomists such as Audouin, Gravier, Quatrefages, Saint-Joseph and Savigny, followed by few more recent ones like Bellan, Bhaud, Gillet, Laubier or Rullier. However, as many parts of the world, the number of taxonomists has dramatically declined in recent years, because taxonomy was not 'sexy' or technology-focused enough to attract policy makers attention. As a result, French scientists were almost absent from the international worm community for the last few decades. Indeed, prior to 2016, almost no French taxonomists participated in the different International Polychaete Conferences, with the exception of the conference organized in Angers in 1992. The absence of French representatives on the council of the International Polychaetology Association meetings was also a reality for several years. Fortunately, French marine biologists are now included in the RESOMAR network, allowing for a new dynamic and the recruitment of several technicians and researchers specialised in identifi cation of benthic fauna. The lack of experienced taxonomists acting as mentors in France was compensated by the motivation of these young scientists. During the past decade, they have published numerous papers on French polychaete taxonomy (e.g., Bonifácio et al. 2015;Jourde et al. 2015;Lavesque et al. 2015Lavesque et al. , 2020cLe Garrec et al. 2017) with the fundamental help of international experts such as Barnich, Blake, Glasby, Hutchings, Meißner, Londoño-Mesa and Parapar among others.

What are the consequences of this hidden biodiversity?
So what difference does it make to know that not only one, but two extremely similar species of Lanice exist? Does this hidden diversity really matter? Of course the answer is yes! Of course, it is essential to use the appropriate name when identifying a species (Lavesque et al. 2019b;Hutchings & Lavesque 2020;Hutchings 2021). Specimens belonging to cryptic or pseudo-cryptic species are very similar and thus diffi cult for people, even for taxonomists, to distinguish. However, as most of these species evolve differently from a common ancestor, their life-traits and their ecological function may be different, or in the process of becoming different. Indeed, in his review on cryptic polychaete diversity, Nygren (2014) shows that many cryptic species can be distinguished by a number of biological characteristics, such as reproductive biology, life history, feeding biology, salinity, habitat and depth preferences or anoxia and temperature tolerances. Each species has a unique set of micro-habitat requirements and functions with important ecological consequences. Misidentifi cation or an underestimation of the diversity thus have a strong impact on ecological studies.
The sand mason worm, Lanice conchilega, is a perfect example to illustrate this point. By aggregating sand particles on its tube, this species acts as ecosystem engineer for forming reef-like structures (Rabaut et al. 2009;Hutchings et al. 2021b). The presence of these biogenic structures increases habitat quality and enhances local biodiversity by changing hydrodynamics and nature of the shore, increasing habitat stability and oxygen supply, and fi nally creating heterogeneity in a uniform environment (Van Hoey et al. 2008). This habitat is thus very attractive for predators like fi shes and foraging waders, and thanks to its high functional value, this habitat also has high conservation value (Godet et al. 2008). By contrast, in Arcachon Bay, the very similar species L. kellyslateri has a scattered distribution with worms appearing solitary. Maybe this absence of a "reef structure" could be linked to a specifi city of this new species or to the particular environment occurring in this lagoon. These worms from Arcachon Bay may not be attractive for birds and perhaps policy makers would be unlikely to protect this species and its habitat. As we can see, the stakes can be high when considering cryptic species individually.
Another example worth considering is the strawberry worm Eupolymnia nebulosa. Experiments conducted on specimens sampled in the Gulf of Lion (Mediterranean Sea) allowed for insights into its feeding mode and tube building (Grémare 1988;Grémare et al. 1989;Grémare & Amouroux 1990), its bioturbation activity (Maire et al. 2007) and development (Bhaud 1988;Bhaud & Grémare 1988). Another study by Grémare (1986) highlighted that two populations, one from Banyuls-sur-Mer (Gulf of Lion), the second from Dinard (English Channel) had different reproductive modes. However, Lavesque et al. (2021) have shown that specimens sampled from Banyuls-sur-Mer belonged to two new species: E. lacazei and Eupolymnia sp. C. Additionally, specimens from Normandy and Brittany (English Channel) belonged to a third new species: E. gili. All these specimens were previously identifi ed as E. nebulosa, but clearly belong to three cryptic species. We can therefore observe that these differences in reproductive modes can be linked to different species rather than different populations, as previously suggested by Martin et al. (1996). Discovery of multiple species with restricted distributions has implications for conservation. For example, it may be assumed that isolated populations can easily recover from local disasters (oil spill for example) by recruitment from nearby populations. But if it turns out that a species previously thought to be widespread is really several different species, this may have implications for recovery from local perturbations.
Regarding non-indigeneous species (NIS), misidentifi cations can have a signifi cant impact on the understanding of ecosystems, and cascading consequences for environmental management if they are detected too late. Frequently, exotic species are often morphologically very similar to native species. As they are not reported from European waters, they are absent from identifi cation keys restricted to the local area. It is therefore essential that the most up-to-date and relevant literature is used to identify species for ecological monitoring studies. Even if this is time consuming and expensive, the species which are typically found in an ecosystem should be regularly checked in detail by using a complete diagnosis, and not just by means of outdated keys that will of course give a poor result. Particular attention should be paid to sensitive areas, where NIS are known to occur, such as harbours, marinas and oyster farms. For example to illustrate this problem, using the blood-worm Marphysa sanguinea (Montagu, 1813) which was largely reported from Arcachon Bay for decades. However, after a thorough morphological and molecular analysis, a second species, new for science was found, i.e., Marphysa victori Lavesque, Daffe, Bonifácio & Hutchings, 2017. This discovery may seem anecdotal, but after extensive investigations, we could confi rm that this species native to South East Asia was probably introduced into the bay via oyster transfers in the 1970's, after mass mortalities of Portuguese oysters (Lavesque et al. 2020c). Moreover, M. victori is an important economic resource as bait and collected both by recreational and professional fi shermen, with about 1 million worms traded per year (Lavesque et al. 2017b), most of them shipped live and sold in Western Mediterranean fi shing shops in France (Lavesque et al. 2020c). Similarly, the presence of the Asiatic terebellid Thelepus japonicus was recently reported for the fi rst time in Europe. Again, its presence in Arcachon Bay and in Normandy is linked to oyster farming with a probable introduction from Japan to Arcachon Bay via oyster transfers and a secondary introduction from Arcachon to Normandy by local transfers (Lavesque et al. 2020a). Prior to the "Spaghetti Project", this species was confused with Thelepus setosus, originally described from France, and therefore absent from European identifi cation keys.
Finally, knowing the exact number of species within a region, or at least the number as close to reality as possible, is fundamental to understanding biodiversity issues. In the context of the extent of the biodiversity crisis, ignoring cryptic species leads to an underestimation of the species richness in the oceans (Bickford et al. 2007;Nygren 2014). Describing this cryptic diversity is absolutely fundamental in the context of the biodiversity crisis (Bickford et al. 2007). We cannot assess the loss of biodiversity in an anthropogenic context if we do not know how many species really occupy an area. Similarly, we cannot identify areas of endemism or areas of biological interest, without knowledge of cryptic species. If we just take into account the results of this "Spaghetti Project", the biodiversity of terebellids has exploded recently, with 31 new species for French waters. Of course, we know that all these new species are not really new but have just been overlooked for ages, representing a hidden biodiversity. The alarming message of how much biodiversity has been underestimated must be clearly conveyed to the public, the politicians and the managers. In the same way, we need to know exactly which species live in an ecosystem to evaluate the effects of global change. For example, recent studies tend to prove a "tropicalization" of the Bay of Biscay, with several species belonging to different biological groups (algae, fi shes, decapods, molluscs or worms) shifting their northern distribution limit from tropical regions north to the southern part of the Bay of Biscay (Portugal, Spain and South of France) (Lima et al. 2007;Arias & Crocetta 2016;Encarnação et al. 2019;Schäfer et al. 2019). Among these species, at least one of them could become problematic. Indeed, the bearded fi reworm Hermodice carunculata (Pallas, 1766), originated from the West Indies and recently observed in southwestern Iberian Peninsula (Encarnação et al. 2019), can cause severe pain if its stinging chaetae come into contact with human skin.

What remains to be done with these Spaghetti worms?
We, as taxonomists, have the responsibility to share our studies and make sure that our work reaches a wide audience. Scientifi c papers and international conferences are not suffi cient and we should use a variety of media (TV, newspapers, social media and blogs) to communicate our fi ndings (Hutchings 2020). Biodiversity is not restricted to geeks of taxonomy and our mission is to help students, ecologists and other professionals to put the right name on the right animal (Hutchings & Lavesque 2020). We also have to explain to politicians why taxonomy is important to the economy and biodiversity conservation, especially with regards to zoning plans for marine parks or management of marine pests (Hutchings 2020(Hutchings , 2021. We have to produce easy to use identifi cation keys, which allows people to differentiate species from cryptic complexes if possible. Our keys should be web-based and thus widely available to the wider biological community (Hutchings et al. 2021b). We, as experts, should be available to help people to identify or confi rm their identifi cations, especially if those seeking for help come from countries lacking taxonomists and/or accurate literature.
Even if our "Spaghetti Project" permitted the improvement of the knowledge of terebellids from French waters, there is still a lot to be done. Firstly, most of the specimens examined were collected in French coastal waters, with the exception of some worms sampled from the deep Capbreton canyon (Lavesque et al. 2019a). The exploration of off-shore deep sea areas should be enhanced in order to have a better understanding of the distribution of these species. Some regions, like the eastern part of the French Mediterranean Sea, were poorly surveyed due to absence of benthic ecologists and samples (i.e., Marseille and Villefranche-sur-mer). This project has highlighted the presence of at least another eight undescribed species in France, based on molecular results. These "orphan" sequences belonged to small or damaged specimens, which were not in good enough condition to be described morphologically. Nygren et al. (2018) obtained similar results while working on Terebellides from Northern Europe; they obtained sequences belonging to 14 still undescribed species (Parapar et al. 2020a). Many more new species probably occur in other parts of Europe where this group was not really studied in detail before, for example in the UK or Italy to name but a few. As discussed before, due to species having restricted distributions (Nygren et al. 2018), more local studies are needed to give us a better picture of the true biodiversity of the region. Globally, some regions like Australia and Brazil are relatively well studied, leading to descriptions of tens of species (Hutchings et al. 2021b), but several regions in the world represent a "taxonomic desert" for terebellids like African, Indian and polar regions (Hutchings et al. 2021b;Capa & Hutchings 2021). So this "Spaghetti Project" could provide a blue print for what is needed in other parts of the world.
For the stability of taxonomic nomenclature, it is important to erect neotypes for old European species described by early naturalists. Indeed, most of these species were only subsequently designated as type species of genera, and very often type specimens were not designated or do not exist anymore and original descriptions are very brief according to current standards. During this project, we highlighted this need for several species like Trichobranchus glacialis and Octobranchus lingulatus, both type species of their genera (Lavesque et al. 2019a), Polycirrus denticulatus (Lavesque et al. 2020b), Amphitrite edwardsii and A. fi gulus, Eupolymnia nebulosa and E. nesidenis (the type species of the genus), Lanice conchilega (the type species of the genus), and fi nally Pista cristata (the type species of the genus) which is currently being redescribed (Londoño-Mesa et al. in prep.;Lavesque et al. 2021) (Fig. 9; Table 1). Obtaining molecular sequences from neotypes is also crucial for future comparison and integrative taxonomy. This ensures that every species will have a modern description based on morphological and molecular tools. Undoubtedly, fi xing neotypes will allow taxonomists to describe new species, as they will have a reference point for comparison. When Parapar & Hutchings (2014) designated a neotype for T. stroemii, they opened the door to the description of 13 new species of Terebellides from Europe, with most of these new species identifi ed in the past as T. stroemii (Lavesque et al. 2019a;Parapar et al. 2020a).
To conclude, the collaborative "Spaghetti Project", supported by numerous enthusiastic people was a real success story. We are aware that some areas and habitats along the French coast are underrepresented in this study but nonetheless, we are sure that it will facilitate the discovery of additional undescribed species not only in our region, but also in the rest of Europe. This focus on the hidden biodiversity of terebellids can be translated to other parts of the world and also to other families, the estimated number of remaining new polychaetes species being greater than 20 000 (Pamungkas et al. 2019;Capa & Hutchings 2021;Magalhães et al. 2021). An interesting challenge will now be to develop online user-friendly tools, like the Delta (Coleman et al. 2010) or Xper (Ung et al. 2010) identifi cation keys. A new volume of the Fauna Iberica collection with a focus on terebellids is also in preparation and coordinated by Julio Parapar. Finally, a "European Terebellids Tour" to sample and erect neotypes of old species should be planned ( Fig. 9)! study was partially funded by the Biodiversity Platform (EPOC laboratory, Arcachon) and generously supported by the Australian Museum, Sydney. Nicolas Lavesque and Guillemine Daffe have received fi nancial support from the French State in the frame of the "Investments for the future" Programme IdEx Bordeaux, reference ANR-10-IDEX-03-02. JMMN receives a productivity grant from Conselho Nacional de Desenvolvimento Científi co e Tecnológico (CNPq), level 2, Brazil. Fig. 9. Type localities of European species for which a neotype is required.