Taxonomical diversity and palaeobiogeographical affinity of belemnites from the Pliensbachian–Toarcian GSSP (Lusitanian Basin, Portugal)

High‐resolution analysis of the late Pliensbachian – early Toarcian belemnite assemblages from the Peniche section (Lusitanian Basin) has enabled, for the first time, recognition of eight taxa of the suborder Belemnitina, previously reported from contemporaneous north‐west Tethyan and Arctic sections. The presence of Bairstowius amaliae sp. nov. in the late Pliensbachian (emaciatum Zone) represents a novelty given that hitherto the genus Bairstowius was known only from late Sinemurian and early Pliensbachian deposits. Additionally, the replacement of Bairstowius amaliae by Catateuthis longiforma, during the latest Pliensbachian, suggests an evolutionary relationship between the two taxa. This relationship suggests a new scenario for the subsequent development of endemic Toarcian Boreal–Arctic faunas, characterized by the occurrence of Catateuthis. Comparison of the Peniche belemnite fauna with coeval faunas from the Mediterranean/Submediterranean and Euro‐Boreal domains indicates taxonomic uniformity during the late Pliensbachian and early Toarcian (emaciatum and polymorphum Zones), in the north‐west Tethys. Despite the lack of a marked taxonomic turnover, the Pliensbachian–Toarcian boundary corresponds to a slight decrease in diversity observed not only in the Lusitanian Basin but also in coeval north‐western European basins. Ordination and cluster analyses indicate that the largest changes in belemnite diversity and palaeogeographical distribution occurred rather during the Toarcian Oceanic Anoxic Event (base of the levisoni Zone). This event is marked by the extinction of taxa, affecting more severely the Mediterranean/Submediterranean domain and resulting in a more pronounced provincial differentiation among north‐western European and Arctic belemnite faunas.

probably related to major palaeoenvironmental and palaeogeographical changes during the Toarcian. However, the study of belemnite diversity during this time-interval has been hampered by a variety of factors, including the poor stratigraphic representation of the Pliensbachian-Toarcian boundary in many European sections (Morard et al. 2003;Pinard et al. 2014), and the lack of high-resolution (ammonoid subzone) analysis of the belemnite assemblages in particular localities (Choffat 1880;Riegraf 1980;Mouterde et al. 1983;Schlegelmilch 1998). Additionally, many studies on Toarcian belemnites are focused on the Euro-Boreal basins (Riegraf et al. 1984;Doyle 1990Doyle , 1992Little 1995;Schlegelmilch 1998;Morten & Twitchett 2009;Caswell & Coe 2014;Xu et al. 2018), whereas little is known from the Mediterranean/ Submediterranean domain (Lissajous 1927;Comb emorel in Rulleau et al. 1998;Sanders et al. 2015;Weis et al. 2015). Therefore, palaeobiogeographical patterns, diversification patterns and evolutionary trends of Early Jurassic belemnites, in the north-west Tethys, remain poorly constrained and some aspects of their provincialism remain speculative (Doyle 1994).
During the Pliensbachian, the north-west Tethys was characterized by a homogeneous European belemnite fauna (Doyle et al. 1994;Weis & Thuy 2015). However, the early Toarcian, characterized by palaeoenvironmental changes and the second-order mass extinction (Toarcian Oceanic Anoxic Event, T-OAE), represents an important period of taxonomic and geographical changes in belemnite assemblages, leading to a Boreal/Tethyan provincialism (Doyle 1994;Dera et al. 2016). In the Euro-Boreal basins, belemnites recovered and experienced a radiation during the middle-late Toarcian (Dera et al. 2016), despite the effects of the T-OAE during the early Toarcian, which caused a reduction in belemnite abundance (Caswell & Coe 2014). The belemnite stratigraphic record in the Mediterranean/Submediterranean domain basins (Italy, Portugal and Morocco) displays a severe reduction in belemnite abundance, or total absence of belemnites, during the T-OAE (latest polymorphumearly levisoni Zones; Sanders et al. 2015;Weis et al. 2015;Ait-Itto et al. 2017;Rita et al. 2019). Moreover, the Toarcian marks the beginning of the endemism of Arctic faunas, probably related to northward migrations of north-west Tethyan groups (Doyle 1987;Doyle et al. 1994), which survived regionally during the early Toarcian crisis, and their rapid evolution into new endemic genera (Sachs & Nalnjaeva 1975;Meledina et al. 2005;Dzyuba et al. 2015).
In this study we present for the first time a detailed systematic description of the Lusitanian Basin belemnite fauna from the late Pliensbachian to the early Toarcian, and a diversity analysis. This study is based on a highresolution stratigraphic analysis of more than 900 specimens collected in the Toarcian GSSP Peniche section. The excellent outcrop conditions, and the abundant belemnite fauna of the Peniche section, enable, on the one hand, a detailed analysis of intraspecific and ontogenetic variation of belemnite taxa in individual samples and, on the other hand, a high-resolution (subzone) study of the belemnite diversity during the late Pliensbachian and early Toarcian. The diversity analysis of the Iberian margin belemnite fauna allows, for the first time, a comparison of diversity patterns at an ammonite zone and subzone scale across the Mediterranean/Submediterranean and Euro-Boreal domains. This comparative approach permits an assessment of the impact of the Pliensbachian-Toarcian crisis on belemnite diversity and a better understanding of the palaeobiogeographical belemnite dynamics during the Early Jurassic of the north-west Tethys.

GEOLOGICAL SETTING
The Peniche section corresponds to the Toarcian GSSP (Rocha et al. 2016) and it is well constrained in terms of ammonite zonation (Duarte & Soares 2002;Duarte et al. 2018). It is located in the Lusitanian Basin, a Mesozoic sedimentary basin developed in the Iberian Western Margin during the North Atlantic Ocean and Occidental Tethyan opening, as a consequence of the Pangea fragmentation (Thierry et al. 2000;Kullberg et al. 2013). The first sedimentary cycle (Wilson et al. 1989) corresponds to extension and rifting episodes and occurred during the Late Triassic -Middle Jurassic (Callovian). It led to the deposition, among others, of a marly limestone succession related to the large opening of the carbonate ramp to the marine environment (Duarte & Soares 2002;Azerêdo et al. 2003;Azerêdo et al. 2014).
The studied section corresponds to a marly limestone succession, 45 m thick, assigned to the upper Pliensbachianlower Toarcian interval, corresponding to an outer ramp environment in an epicontinental sea (Duarte & Soares 2002;Duarte 2007;Duarte et al. 2018; Fig. 1). The end of the Pliensbachian, represented by the Lemede Formation, corresponds to an alternation of decimetric marly limestones and centimetric marls, 11.2 m thick. The next c. 11 m of the succession, belonging to the Toarcian (polymorphum Zone), corresponds to the first member of the Cabo Carvoeiro Formation (CC1) and consists of an alternation of bioturbated marls, with marly limestones (Fig. 1). The majority of this member is highly fossiliferous, containing, among others, belemnites, pyritized ammonites, brachiopods and bivalves (Duarte & F I G . 1 . Stratigraphic distribution and absolute abundance of belemnite species (per bed and per m 2 ) in the Peniche section (Lusitanian Basin, Portugal). Fewer than two occurrences per bed are not represented (see Rita et al. 2020, tables S1-S2 for the full dataset). The Pliensbachian-Toarcian boundary event and the Toarcian Oceanic Anoxic Event (T-OAE) are highlighted and were positioned based on Rocha et al. (2016) and Duarte et al. (2018). Sequence stratigraphy according to Duarte (2007). The bed numbers in square brackets correspond to merged beds due to sample size constraints, regarding the Peniche diversity analysis, and follow Rita et al. (2019). Stratigraphic log based on Rita et al. (2019). Ammonite zonation and subzonation from Mouterde (1955) and Rocha et al. (2016), respectively. Lithostratigraphy from Hesselbo et al. (2007). Bed numbers from Rita et al. (2019). Abbreviations: BG, belemnite gap; elisa, elisa/hawskerense Subzone; mir., mirabile/paltum Subzone; solare, solare/apyrenum Subzone.
Soares 2002). The upper 25 m of the studied succession belongs to the second member of the Cabo Carvoeiro Formation (CC2, levisoni Zone; Fig. 1). This part of the succession is represented by siliciclastic-rich marls, interbedded with sandy marly limestones and rare carbonated sandstones and microconglomerates, clearly associated with turbiditic deposition (see Wright & Wilson 1984;Duarte 1997). For that reason, the occurrence of benthic macrofauna in these horizons is rare (some horizons with brachiopods), and both benthic and planktonic organisms become scarcer upwards in the sedimentary succession (Comas-Rengifo et al. 2015;Rita et al. 2016;Correia et al. 2017). Ammonites occur in the whole CC2 member. The levisoni Zone corresponds to a barren interval for belemnites (Rita et al. 2019), except for one single horizon, where 11 poorly preserved specimens were found (only two specimens could be identified to the genus level; Fig. 1). This level is not well-defined biostratigraphically at the subzone scale, but recent strontium isotopic data on belemnites and brachiopods seem to indicate the uppermost exaratum Subzone (= elegantulum/levisoni; McArthur et al. 2020) rather than the base of the falciferum Subzone. This means that belemnites are missing for almost an entire subzone in the Peniche section. Figure 2 summarizes the ammonoid zones and subzones covered by the present study and how they correlate with biostratigraphical schemes from coeval oceanic domains. As indicated in Figure 1, in the solare/apyrenum Subzone, beds P1 and P2 were sampled; in the elisa/ hawskerense Subzone, beds P3a, P3b and P4 were sampled; in the mirabile/paltum Subzone, bed P5 was sampled; in the semicelatum Subzone, beds P6, P7, P8, P9a, P9b and P9c were sampled; and in the elegantulum/levisoni Subzone, bed 10 was sampled.
We focused on quantitatively collecting belemnites from well-exposed bedding planes. The belemnite rostra were collected by (1) sampling all specimens in 14 1 m 2 areas; and (2) collecting at least 30 complete specimens (with at least the alveolus preserved) in the remaining bed area, i.e. outside of the quadrats, when possible. This first sampling method allowed the calculation of absolute abundance, considering both fragmented and complete specimens. The whole dataset, regardless of the sampling method, was used for the taxonomic and diversity analyses and for the calculation of the relative abundance of taxa.
Prior to identification, the collected specimens were mechanically prepared with a sand blaster and an air scribe. The photographed specimens were coated with heated ammonium chloride powder. Belemnite species identification was based on the analysis of traditional features, such as shape (outline and profile) and the presence of grooves in the apical region. The transverse section, the depth of penetration of the phragmocone and the apical line were observed on micro-computed tomography (lCT) using our in-house lCT phoenix v|tome|x s 240 (Research Edition) scanner (Rita et al. 2019(Rita et al. , 2020) but also in longitudinally cut and polished specimens. All these features were afterwards compared with published descriptions and figures (Riegraf et al. 1984;Doyle 1990Doyle , 1992Schlegelmilch 1998;Pinard et al. 2014;Sanders et al. 2015;Weis et al. 2018). This method allowed us to recognize the features of each ontogenetic stage, due to the observation of epirostra and growth lines. It was possible to distinguish between adult (ephebic-gerontic sensu Doyle 1990), neanic (sensu Doyle 1990) and juvenile (nepionic sensu Doyle 1990) specimens.
Systematic descriptions include size measurements of well-preserved specimens, taking into account the following metrics: total preserved length (L); length of the rostrum solidum (apical length, l); alveolar angle (AA); dorsoventral diameter (Dv), lateral diameter (Dl) at the protoconch level; cross-sectional distance from protoconch to ventral side (Rv) or to dorsal side (Rd); and width (W) and height (H) at the aperture level ( Fig. 3), following the terminology of Doyle (1990) and Dera et al. (2016). A total of 277 specimens were CTscanned to allow the acquisition of all the metrics, given that the alveolus was filled with sediment. In addition, 654 specimens were measured with a digital calliper. The rostrum shape was assessed by calculating the robustness (L/H), compression (Dv/Dl) and elongation (Dv/l) indices. The robustness index (Rob.) discriminates stocky (c. 2), robust (c. 2-c. 10), or slender specimens (>10). The compression index (CI) indicates the general shape of the alveolar aperture and/or crosssections. The calculated values may be lower than, equal to or greater than 1, and refer to depressed, regular, or compressed shapes, respectively. The elongation index (EI) discriminates elongated (higher values) from short forms (lower values). The eccentricity (E = Rd À Rv/DvÁ100%) of the protoconch indicates if the protoconch is deviated towards the ventral or to the dorsal side of the rostrum.
The size categories used in systematic descriptions were previously defined by Weis et al. (2018) and refer to the maximum total preserved length of the rostrum (L) of the adult specimens (when ontogenetic stages were possible to identify) as follows: very small (<30 mm), small (30-60 mm), medium (61-100 mm) and large (101-150 mm; Fig. 4).

Diversity analysis
For the diversity analysis of the Peniche data the species richness (no. species; S), rarefied species richness (sample size-based rarefaction) and Shannon-Wiener index (Shannon & Weaver 1949) were calculated. The data were analysed at the bed scale but, due to sample size constraints (e.g. poor outcrop conditions), the data from beds P9b and P9c were merged. Sample P10 was not considered for the rarefied species richness because only two specimens were able to be identified at the genus level. Rarefied species richness was calculated using the function rarefy from the vegan package (Oksanen et al. 2018) in R (R Core Team 2018), which corrects for the sample size by using a subsampling method. Shannon-Wiener diversity index was calculated with the function diversity. Passaloteuthis sp. juv. was not included in this analysis due to the presence of P. milleri and P. bisulcata in the levels where it is represented. Hastitidae sp. indet. and Acrocoelites sp. indet. were included in the analysis.
In order to compare belemnite diversity of the Lusitanian Basin with coeval Tethyan basins, ammonite zone and subzone scales were used. When abundance data were not available to allow rarefied species richness to be calculated, presence/absence data were used and raw species diversity S was calculated.
Belemnite fauna from Peniche were compared with data from: (1) Fig. 5). It should be noted that only specimens classified to the species level (except for Hastitidae indet. in Lusitanian and Asturian basins) were included in the diversity analysis and only ammonite zones with 20 or more specimens were considered.
The published data on the belemnite assemblages from the Swabo-Franconian Basin (Schlegelmilch 1998; Riegraf F I G . 3 . Longitudinal view (A) and transverse cross-section of a belemnite at the protoconch (B) and aperture (C) levels of a belemnite rostrum, with the measured parameters indicated: alveolar angle (AA), dorsoventral diameter (Dv), lateral diameter (Dl), height (H), apical length (l), total preserved length (L), cross-sectional distance from protoconch to ventral side (Rv) and cross-sectional distance from protoconch to dorsal side (Rd) and width (W). et al. 1984) at the zone and subzone level do not include highly resolved abundance data. Additionally, South Riffian Basin data (Morocco; Sanders et al. 2015) are limited to the polymorphum Zone (lower Toarcian). Therefore, presence/absence data were used to perform the nonmetric multidimensional scaling (nMDS) based on Bray-Curtis distances and cluster analyses, in order to allow a comparison between all the north-west Tethyan belemnite faunas described as hitherto belonging to the late Pliensbachianearly Toarcian (emaciatum-levisoni zones) interval. Analysis of similarities (ANOSIM) was used to determine whether the different clusters identified in the nMDS analysis differed significantly (Heaven & Scrosati 2008). This was done using the function anosim from the vegan package (Oksanen et al. 2018) in R (R Core Team 2018).
In this study, we considered the Lusitanian and South Riffian basins as part of the Mediterranean/Submediterranean domain and the remaining studied basins (Cleveland, Asturian, Causses and Swabo-Franconian) as part of the Euro-Boreal domain, except for Russia (northern Siberia and Russian Far East), which is included in the

Remarks
Remarks. The genus Bairstowius was originally included in Hastitidae family rather than Passaloteuthididae, based on the 'Hastites-like pattern of lateral furrows, its compressed rostrum, and absence of apical grooves ' (Doyle et al. 1994, p. 12). Subsequent publications (Doyle 1994(Doyle , 2003(Doyle , 2010Riegraf et al. 1998, 'Subhastitidae';Schlegelmilch 1998) maintained this classification without further discussion. Further morphological analysis as demonstrated by Bolton (1982) and hereafter (Peniche) call for a slight emendation of the original diagnosis. In fact, short dorsolateral apical grooves are commonly developed in both Bairstowius longissimus (Bolton 1982) and B. amaliae, although they always remain weak, compared with Passaloteuthididae. The presence of weakly developed dorsolateral grooves in some individuals can be interpreted as an atavistic character, given that Bairstowius is commonly considered the phylogenetical link between passaloteuthidids and hastitids (Schwegler 1962;Schumann 1974;Bolton 1982). The development of an epirostrum has been noted by Bolton (1982) and it is confirmed by our material. Epirostra are not known for the genera Subhastites, Hastites, Rhabdobelus, Parahastites or Sachsibelus but are commonly developed in Pleurobelus, a genus included into Hastitidae by some authors (Jeletzky 1966;Doyle 1994;Pinard et al. 2014). However, according to Doyle (1990, p. 14), the presence or absence of an epirostrum cannot be considered as a valuable taxonomic criterion for differentiation of genera. Consequently, the presence/absence of an epirostrum is not considered here a diagnostic character retained at the genus level.
Occurrence. Upper Sinemurianupper Pliensbachian of England, France, Germany, Italy and Portugal. mediolateral line weakly developed in the stem region only; dorsolateral line extending from the apical to the stem region, fading out on the alveolar region ( Fig. 7). Epirostrum commonly developed and the apical line ortholineate. The alveolus occupies c. one-quarter to one-fifth of the orthorostrum.
Differential diagnosis. The morphology of B. amaliae can be placed between the stouter Bairstowius charmouthensis (Mayer 1864) and the more gracile elongate B. junceus (Phillips 1867) and B. longissimus (Miller 1826). The latter taxa are stratigraphically older than B. amaliae (early Pliensbachian, jamesoni and ibex Zones). B. junceus and B. longissimus differ by their longer, more elongated rostra, the cylindrical profile and the subcircular cross-section. B. charmouthensis is the species that most resembles B. amaliae by its hastate to subhastate profile and outline, and its compressed cross-section. However, it differs by the absence of epirostrum and dorsolateral apical grooves, and by a deeper penetration of the alveolus (c. one-third of total rostrum). B. amaliae is morphologically and stratigraphically replaced by Catateuthis longiforma at the Pliensbachian-Toarcian boundary in the Peniche section. The species C. longiforma differs from B. amaliae by a cylindriconical outline and profile, a more eccentric protoconch, a different pattern of lateral lines (Fig. 7), a stouter (lower Rob.) and less elongated rostrum, and a higher compression of the rostrum (Fig. 8).
Description. The small-to-medium-sized, spicular elongated rostrum shows an asymmetrical hastate profile and a symmetrical subhastate to cylindrical outline. The apex is needle-like and some specimens bear two very weak dorsolateral apical grooves in the continuation of the lateral lines. This is, however, not a stable feature. A pattern of three parallel lateral lines is observed ( Fig. 7A): in a ventrolateral position a deep, broad and well-defined groove can be observed; this main lateral line/groove has a 'Doppellinien' pattern. It is separated by a weak ridge from the lesser incised second lateral line, which is developed as a shallow depression only in the stem region, in a mediolateral position. The third lateral line is positioned in a dorsolateral position, extending from the apex to the alveolar region, fading out on the latter (Fig. 7A). The crosssection is compressed, being pyriform in the apical region and quadrate in the stem and alveolar areas (Figs 7A, 9A). The dorsal alveolar area is characterized by a flattening. The phragmocone occupies one-quarter to one-fifth of the orthorostrum. The alveolar angle varies between c. 19°and 23°(Rita et al. 2020, table S7). A cylindriconical epirostrum is sometimes developed in the larger specimens (8 specimens out of a total of 107), in combination with an attenuated and striated apex. The cross-section of the epirostrum is subcircular to slightly pyriform. The transition between epirostrum and orthorostrum is extremely gradual from an external point of view and can be readily observed only in longitudinally sectioned specimens (Fig. 6I). Remarks. The small size of the rostrum and the poorly developed diagnostic characters could indicate that these specimens represent juveniles of Hastitidae. It is, however, not possible to assign these specimens to a genus with certainty, due to the lack of diagnostic characters. We can exclude the juvenile forms of Bairstowius amaliae or Catateuthis longiforma because they have a blunt apex. Additionally, the stratigraphic distribution of Hastitidae sp. indet. differs from the range of B. amaliae or C. longiforma (Fig. 1).  mat., figs S1 B1-B3.
Description. Small to medium-sized slender forms with an acute apex bent towards the dorsal side. The outline is elongated, cylindriconical and symmetrical, the profile differing only by its slight asymmetry. The profile can be subhastate, especially in the juvenile forms. Two dorsolateral grooves can be observed: they are short and weakly developed in the juvenile and neanic specimens, but become more pronounced in the adults. A pattern of two lateral lines is observed. The upper one, located in the dorsolateral area, extends from the stem to the apex with a variable length and is connected to the apical dorsolateral groove. The lower lateral line, located in the ventrolateral area, is better developed in the alveolar region, extending to the stem. Rarely, a dorsal groove in the alveolar region (probably pathological) can be observed. The cross-section is compressed, circular/subcircular in the apical region and quadrate in the alveolar region (Fig. 7B). A flat area is developed in the dorsal part. The phragmocone occupies one-third to one-quarter of the orthorostrum and the apical line is ortholineate. A depressed section can occur in some individuals (Rita et al. 2020, table S9;specimen 2018.10.136). An epirostrum is sometimes developed (47 out of 359 specimens) in larger specimens (L = 23.4-67.7 mm). The boundary between epirostrum and orthorostrum is visible only in longitudinal section (Fig. 6J). The epirostrum is elongated and acute, bearing abundant striae. If the phragmocone is not considered, the solid orthorostrum/epirostrum ratio is 1.82-1.86 (Rita et al. 2020, table S9).
Remarks. The specimens from Peniche differ from those described in the literature (Doyle 1990;Sanders et al. 2015;Weis et al. 2015Weis et al. , 2018 by the occasional development of an epirostrum. Additionally, the specimens described herein are slightly more slender and more elongated than those described by Doyle (1990; Fig. 11). These differences are, however, insufficient to consider these specimens as a new taxon. The single specimen shown in longitudinal section by Doyle (1990, pl. 3, fig. 9) does not show an epirostrum, but its size and ontogenetical development indicate that it might not be a fully grown specimen, and the epirostral development is considered a feature of the later (adult) growth stages (M€ uller-Stoll 1936;Arkhipkin et al. 2015). Peter Doyle (pers. comm. 2019) stated that no epirostra were found within the set of specimens from the Cleveland Basin he analysed. Combining the morphological and stratigraphic information from Peniche with literature data, we outline the possibility of a lineage Bairstowius-Catateuthis, with Bairstowius amaliae and Catateuthis longiforma as connecting taxa between both genera. This morphological affinity had already been acknowledged by Doyle (2003), who included Belemnites longiformis into Bairstowius genus.
Occurrence. Lower Toarcian (polymorphum Zone) of the South Riffian Basin of Morocco, upper Pliensbachianlowermost Toarcian (tenuicostatum = polymorphum Zone) of the UK and mainland Europe. Description. Very small to small rostra with a characteristic short apical region. The profile is asymmetrical and cylindriconical, the outline symmetrical, cylindrical to subhastate. The apex is mucronate and bent towards the dorsal side. The development of apical striae is common. Two weakly developed, short dorsolateral grooves are observed. Two lateral lines, represented as shallow depressions, separated by a weak ridge, can be recognized: a medio-dorsolateral line (broader and better developed) and a ventrolateral line (weakly developed). Both extend from the alveolar region to the stem, not reaching the apical region. The cross-section is subcircular to slightly pyriform in the alveolar region and subquadrate in the apical region, due to the lateral depressions. The phragmocone penetrates one-half to one-third of the rostrum and is slightly ventrally displaced. The apical line is goniolineate to slightly cyrtolineate. Alveolar angle varies between c. 21°and 24°(Rita et al. 2020, table S10). A small epirostrum can be seen (Fig. 13).
Description. The medium to large-sized robust specimens show a symmetrical and conical to cylindriconical profile and outline. The apex is acute and bears two dorsolateral grooves weakly developed in most individuals. The cross-section is circular to pyriform in the apical region and subquadrate in the stem and alveolar regions, sometimes slightly compressed. The phragmocone is slightly eccentric, penetrating one-third of the rostrum, except for some specimens, where higher values of penetration can be observed. The apical line is goniolineate. The alveolar angle ranges between c. 20°and 24°(Rita et al. 2020, table S11).
Ontogeny. The CI and Rob. increase during ontogeny while the EI decreases (Fig. 9D).
Description. The medium-sized, elongated rostrum is characterized by a cylindrical to cylindriconical profile. The outline is cylindrical to subhastate. The apex is moderately acute, occasionally with striae developed only in the ventral area; dorsolateral grooves are developed as well-defined depressions. The cross-section is slightly compressed, subquadrate to subcircular in the alveolar region but pyriform in the apical area. A lateral flattening is present in the alveolar region. The apical line is ortholineate to slightly goniolineate, and the phragmocone occupies one-quarter to one-third of the rostrum (Rita et al. 2020, table S12).
Description. Very small to small rostra with a conical symmetrical profile and outline and acute apex. The cross-section is quadrate to circular in the alveolar region but pyriform in the apical region. The phragmocone penetrates one-half to one-third of the rostrum (Rita et al. 2020, table S13). The apical line is ortholineate or slightly goniolineate.
Remarks. These specimens can be assigned only as juveniles of Passaloteuthis due to the main features described above. However, the absence of other taxonomically relevant features such as lateral lines and apical grooves, which usually develop later in ontogeny, does not allow a species-level classification. These specimens are, however, probably juvenile forms of the most common taxa P. bisulcata or the rarer P. milleri.
Description. The specimen MCUC 2018.10.002 corresponds to a large, slender rostra with a cylindriconical symmetrical profile and outline. The apical region is very elongated, and the apex is acute. The cross-section is subcircular in the alveolar region but pyriform in the apical region. The phragmocone penetrates onethird of the rostrum (Rita et al. 2020, table S14). A ventral apical groove occupies one-third of the whole rostrum and two weak dorsolateral grooves are observed in the apical region. The apex bears long striae. The specimen MCUC 2018.10.001 is a large, stout rostrum with a conical symmetrical profile and acute apex.
Remarks. The specimens were embedded in calciclastic sediment, which precluded a non-destructive preparation process. The preparation allowed only one side to be observed in one of the specimens (MCUC 2018.10.001) and the surface of both specimens could not be properly cleaned. Therefore, several taxonomically relevant features could not be measured or analysed, such as penetration of the phragmocone, apical line, cross-section and the presence of grooves and lateral lines. The absence of these taxonomically relevant features hampers a species-level classification. Other authors also report only Acrocoelites sp. indet. from these beds (McArthur et al. 2020).
Biodiversity patterns. At the bed scale, belemnite diversity is not constant during the late Pliensbachianearly Toarcian interval in the Peniche section, according to S, Shannon-Wiener index and rarefied species richness (Fig. 15). Due to the observed sample size oscillations (ranging from 2 to 174 in 13 beds, see Rita et al. 2020, table S1), rarefied species diversity might be the more reliable measurement of the diversity in the Peniche belemnite fauna. This approach standardizes samples by size, drawing down samples to equal numbers of specimens, normalizing richness to a standard sample size (Sanders 1968;Zhao et al. 2014). This is particularly important due to the nature of the sedimentary record, which does not allow a uniform sampling because of facies preservation, outcrop conditions or taphonomy (preservation probability), for example Smith & McGowan (2011).
An increase in the diversity (S, Shannon-Wiener index and rarefied species richness) is observed during the upper Pliensbachian (emaciatum Zone, beds P1-P3a). The interval ranging from P3a to P8 corresponds to a diversity drop, followed by an increase until the end of the polymorphum Zone (Shannon-Wiener index and rarefied species richness). Diversity dramatically decreases (Shannon-Wiener index and S) from the polymorphum-levisoni zonal boundary (beginning of the T-OAE) until bed P10 (Fig. 15B).
At the subzone scale (Fig. 15B), diversity increases from the solare/apyrenum Subzone to the elisa/hawskerense Subzone, followed by a decrease at the Pliensbachian-Toarcian boundary (rarefied species richness, Shannon-Wiener index and S) in the Lusitanian Basin.
During the T-OAE (latest semicelatumelegantulum/ levisoni Subzone), a decrease in belemnite diversity is observed in the Lusitanian Basin (Shannon-Wiener index and S; Fig. 15B).
In the nMDS and cluster analyses (Fig. 16), each point represents a zonal assemblage from one of the seven sites considered (Asturian Basin, Western Paris Basin, Swabo-Franconian Basin, Cleveland Basin, Lusitanian Basin, Causses Basin and South Riffian Basin). According to the results, two groups can be distinguished in terms of faunal similarities. The first group corresponds to the assemblages ranging from the emaciatum to the polymorphum Zone interval (Asturian, Western Paris, Cleveland, Lusitanian and Swabo-Franconian Basins and Riffian Basin; Fig. 16). The second group consists of all the levisoni Zone (lower Toarcian) belemnite assemblages (Causses, Asturian, Western Paris, Cleveland and Swabo-Franconian Basins; Fig. 16). Within the latter, three clusters can be distinguished: Causses Basin (cluster 1); Asturian and Western Paris Basins (cluster 2); and Swabo-Franconian and Cleveland Basins (cluster 3). The clusters are significantly different according to ANOSIM test, which gives an R value of 0.49 and a p-value of 0.001 (Rita et al. 2020, fig. S1).
It should be noted that the data from the Lusitanian Basin corresponding to the levisoni Zone have not been included in the nMDS analysis due to the lack of specimens determined to the species level (only two Acrocoelites sp. indet. were identified). Although McArthur et al. (2020) have reported ?Pleurobelus sp. A from the polymorphum Zone, our samples did not include this genus. Given that the McArthur et al. specimens were not described or figured nor the abundance recorded, we did not include them in our diversity or comparative analyses.
Biodiversity patterns. At the subzone scale (Fig. 15B), during the late Pliensbachian, diversity increases from the solare/apyrenum to the elisa/hawskerense Subzone, followed by a decrease at the Pliensbachian-Toarcian boundary (rarefied species richness, Shannon-Wiener index and S) in the Lusitanian Basin. This trend is comparable with the changes in belemnite diversity observed in the Asturian, Western Paris and Cleveland Basins (Caswell & Coe 2014) at the zone scale (rarefied species richness; Fig. 17A). At the subzone scale it is comparable with Cleveland Basin, where belemnite diversity increases from the apyrenum to the hawskerense Subzone (see Rita et al. 2020, fig. S2).
During the T-OAE (latest polymorphumearly levisoni Zone), a decrease in belemnite diversity is observed in the Lusitanian Basin (Shannon-Wiener index and S; Fig. 15B). This is comparable to the diversity trend observed in the Western Paris (rarefied species richness) at the zone level and in the Asturian (S) Basins at both the zone and subzone scales (Fig. 17). New analyses based on previously published data from the Cleveland and Swabo-Franconian Basins (Riegraf et al. 1984;Doyle & Bennett 1995;Little 1995;see Material and Method), in contrast, reveal an increase in belemnite diversity (S) during the T-OAE at the zone scale (Fig. 17B). For instance, after the T-OAE, in the bifrons Zone the diversity (S) keeps increasing in the Cleveland and Swabo-Franconian Basins, which is also observed in the Causses Basin. In the variabilis Zone, the diversity decreases (S) in the Cleveland and Western Paris Basins (Fig. 17B). At the subzone scale, however, the T-OAE (exaratum Subzone) corresponded to a decrease in diversity (S) for the Swabo-Franconian Basin, while in the Cleveland Basin the diversity (S) increases (Fig. 17C).

Macroevolutionary context of the Early Jurassic belemnite assemblages from Peniche
The interpretation of the diversity and of the palaeogeographical distribution patterns of Early Jurassic belemnites in the Tethys Ocean is restricted due to the availability of highly resolved biostratigraphy data, which are biased towards the Euro-Boreal domain. This is related to the lack of belemnite collections resolved to the ammonite . Nonetheless, some progress has been achieved and general outlines of diversity and palaeogeographical distribution may be expanded (Doyle 1987;Challinor 1991Challinor , 1992Doyle et al. 1994;Iba et al. 2012;Pinard et al. 2014;Sanders et al. 2015;Weis & Thuy 2015;Weis et al. 2018).
Despite the similarities between European belemnite faunas during the late Pliensbachian and earliest Toarcian, the presence of the new species B. amaliae in the upper Pliensbachian of Peniche makes the Lusitanian Basin belemnite fauna unique. Bairstowius has been hitherto identified only in the upper Sinemurianlower Pliensbachian of England, Germany, France, Spain, western Turkey, andcentral Italy (Doyle 1994, 2010;Schlegelmilch 1998;Weis et al. 2015). Therefore, the record of B. amaliae in the Lusitanian Basin extends the stratigraphic range of the genus Bairstowius to the upper Pliensbachian (emaciatum Zone) in the north-western Tethys. Additionally, it expands the biogeographical range of the Bairstowius genus to the Lusitanian Basin.
Another unique feature of the Lusitanian Basin belemnite fauna is the high relative abundance of the species C. longiforma in the assemblage. This taxon was previously identified in Morocco (South Riffian Basin;Sanders et al. 2015) and in the Cleveland Basin (Caswell & Coe 2014) where it comprised 1-10% of the assemblage, while in Peniche it comprises 30-65% of the assemblage (Fig. 18). This indicates that this taxon might have a preference for particular environmental conditions present in the Lusitanian Basin, but its exact environmental preferences are currently hard to elucidate. Rita et al. (2019) found that this species is particularly sensitive, in terms of body size, during the Pliensbachian-Toarcian warming event coinciding with a rise in surface seawater temperature, deoxygenation, input of nutrients and various other biotic and abiotic changes.
The stratigraphic replacement of B. amaliae by C. longiforma in the Peniche assemblage during the uppermost Pliensbachian, together with the morphological features shared by the two taxa (see Systematic Palaeontology section), support the possibility of a lineage Bairstowius-Catateuthis, with B. amaliae and C. longiforma connecting both genera (see also Doyle 2003, who included Belemnites longiformis into Bairstowius genus for this reason). The possible phylogenetic relation between B. amaliae and C. longiforma offers an evolutionary scenario for the European lowermost Toarcian fauna and the endemic Boreal-Arctic fauna that developed during the falciferum Zone in northern Siberia (Dzyuba et al. 2015), corroborating the hypothesis of a northward migration during the lower Toarcian from the Mediterranean/ Submediterranean domain to the Euro-Boreal domain.

Belemnite diversity patterns across the north-western Tethys
The Early Jurassic has been considered a major bottleneck in belemnite evolution as reflected in the diversity decline from the lower-middle Toarcian (Dera et al. 2016). However, the relative contribution of the perturbations at the Pliensbachian-Toarcian boundary event and the T-OAE is often difficult to assess due to lack of availability of highresolution biostratigraphic data for belemnites with regard to the ammonoid subzones.
The belemnite diversity in the Lusitanian Basin slightly decreases across the Pliensbachian-Toarcian boundary. We cannot entirely rule out the effect of preservation on this pattern, given that rare species such as Passaloteuthis milleri temporarily disappear (Lazarus effect; compare with Twitchett 2007). However, our largest samples are reported from the Pliensbachian-Toarcian boundary interval, what might mean that this is probably a genuine pattern, rather than a sampling artefact. Furthermore, this trend is also observed in the Asturian (herein preliminary results), Swabo-Franconian (Schlegelmilch 1998) and Cleveland (Doyle 1992; Caswell & Coe 2014) Basins. The overall decrease from the late Pliensbachian to the early Toarcian is also comparable with data from the Western Paris Basin, where the diversity decreases (rarefied species richness) from the margaritatus Zone (late Pliensbachian) to the polymorphum Zone (early Toarcian), although no data are available for the emaciatum Zone (Weis et al. 2018). In the Causses Basin, the diversity (S) in the margaritatus Zone is much higher than in the serpentinum Zone (= levisoni, early Toarcian). From this basin, no belemnite data are available from the tenuicostatum Zone (= polymorphum, early Toarcian), hampering analysis of the diversity patterns across the Pliensbachian-Toarcian boundary and the T-OAE (Pinard et al. 2014).
The T-OAE, dated from the latest polymorphum to early levisoni Zone (~exaratum Subzone), corresponds to a severe decrease in belemnite diversity in the Asturian, Western Paris and Lusitanian Basins, in contrast with the increase observed in the Cleveland and in the Swabo-Franconian Basins (at the zone scale). It is, however, noteworthy that the belemnite record is absent or rare in many sections during the early Toarcian, especially in the levisoni Zone in the Mediterranean/ Submediterranean domain (Weis & Thuy 2015). In fact, in Peniche, the beginning of the T-OAE (early levisoni Zone) is almost coincident with a gap in the belemnite record, ranging from the polymorphum/levisoni zone boundary to the bonarelli Zone (late Toarcian; Duarte 1997). This gap is interrupted by the sparse record of In contrast, the levisoni Zone in the northern part of the Euro-Boreal domain (Swabo-Franconian and Cleveland Basins), is characterized by an increase in belemnite diversity (S), while various genera and species are not reported in the Mediterranean/Submediterranean domain, such as Youngibelus and Simpsonibelus (Figs 17, 19). However, when diversity (S) is analysed at the subzone scale, this increase seems to be more acute after the T-OAE, in the falciferum Subzone ( Fig. 17; compare with Riegraf et al. 1984 andCaswell &Coe 2014). Additionally, when analysing belemnite data from Cleveland Basin (Caswell & Coe 2014) at the bed scale, a diversity decrease (S) is observed in the aftermath of the T-OAE, in the early falciferum Subzone, followed by an increase in the middle/late falciferum Subzone. This suggests that the high taxonomic turnover in these (sub)zones might mask the T-OAE extinction, at lower stratigraphic resolution, and might also be responsible for the seemingly high diversity in the levisoni Zone, or individual subzones, in these regions. Further high-resolution (bed and/or subzone scale) analysis of standardized species richness is necessary to disentangle the belemnite diversity dynamics and palaeogeographical patterns during the T-OAE, across the north-west Tethys latitudinal gradient.
The results from the nMDS and cluster analyses support the thesis of a belemnite taxonomic uniformity during the latest Pliensbachianearliest Toarcian (emaciatum and polymorphum zones) in Europe and adjoining areas. The largest changes in belemnite diversity and palaeogeographical distribution occurred during the T-OAE (levisoni Zone), rather than at the Pliensbachian-Toarcian boundary in the north-west Tethyan Ocean. These results seem to differ from the response observed in ammonoids, which have larger differences between Mediterranean and Euro-Boreal faunas across the Pliensbachian-Toarcian boundary and more uniform (cosmopolitan) faunas during the T-OAE (at least at the zone level: compare with Dera et al. 2011). Our results also emphasize the marked differences between the northern (Cleveland and Swabo-Franconian Basins) and the southern part (Causses and Asturian Basins) of the Euro-Boreal domain in terms of belemnite taxonomic composition during the levisoni Zone, not recognized before that. This, at first glance, also differs from the diversity pattern observed in ammonoids, characterized by marked changes in diversity at the Pliensbachian-Toarcian boundary (Dera et al. 2010).
Our quantitative analysis is also consistent with Doyle (1994) and Sanders et al. (2015), who suggested that the European (s.l.) early Jurassic belemnite faunas had a similar composition until the Toarcian. Consequently, during Toarcian and Aalenian, significant changes took place, with a high diversification and a trend towards endemic Tethyan and Boreal-Arctic belemnite faunas (Doyle 1994;Weis et al. 2018). This provincialism was thought to have been triggered by the biotic crisis and palaeoenvironmental perturbations occurring during the early Toarcian, particularly the T-OAE (Doyle 1987;Doyle 1994). The results of our diversity analysis support this interpretation due to the minor decreases observed across the Pliensbachian-Toarcian boundary and more major changes across the T-OAE in various north-west Tethyan basins during the lower Toarcian, despite some regional differences. Further research is necessary to corroborate and understand the underlying reasons for the discrepancies in diversity and biogeographical patterns in belemnites compared with ammonoids.
The presence in the emaciatum Zone (= spinatum, upper Pliensbachian) of numerous specimens ascribed to the genus Bairstowius represents a novelty, together with the high abundance of Catateuthis longiforma comb. nov. in younger samples (uppermost emaciatum and polymorphum zones), in comparison with other European sections. Moreover, the replacement of Bairstowius amaliae by Catateuthis longiforma during the uppermost Pliensbachian suggests an evolutionary relationship between the two taxa.
Despite the lack of a marked taxonomic turnover, the Pliensbachian-Toarcian boundary might correspond to one of the pulses of the Pliensbachian-Toarcian crisis by a slight decrease in species diversity observed not only in the Lusitanian Basin but also in coeval basins (Cleveland, Swabo-Franconian, Western Paris, Causses and Asturian Basins). However, the biggest changes in the belemnite fauna in the north-western Tethys are observed during the levisoni Zone, corresponding to extinction of dominant taxa as well as originations. The extinctions are particularly severe in the Mediterranean/Submediterranean domain and in the southern part of the Euro-Boreal domain, and contribute to a provincial differentiation among north-west European and Arctic belemnite faunas. In Peniche, belemnites are largely absent in the levisoni Zone, with the exception of Acrocoelites sp. indet. in a single bed after the T-OAE. Our study highlights that, in order to fully disentangle belemnite diversity and palaeogeographical dynamics across the north-western Tethys latitudinal gradient during the T-OAE, high -resolution abundance data and sample standardized diversity studies (bed or subzone scale) are necessary.