A new Middle Devonian chitinozoan assemblage from northern Iberia of Palaeobotany and Palynology

Chitinozoans were recovered from theNaranco, Huergasand Gustalapiedra formationsof northernSpain,which yielded a diverse assemblage including 29 taxa in 9 named genera. The deposits are of early Givetian age, a time when Iberia was isolated from large continents as part of the Armorican Terrane Assemblage. The formations studiedhereconsistofa largeclasticunitinterruptingcarbonatedeposition,withlargelimestoneformationspo-sitioned above and below the formations analyzed here. This clastic unit includes the Ka č ák Event, an important global extinction event associated with anoxia in the marine realm. In this paper, the relatively well-preserved chitinozoan assemblage is described and considered from a biogeographical and stratigraphical perspective. The chitinozoan community was deposited in a short space of time and includes various taxa not previously knownfromtheMiddleDevonian,thoughtheassemblageasawholeisattributabletotheperiod.Onlymoderate similarityisseenwith assemblages reported fromLaurussiaand Gondwana,with a slightbiastowardsthelatter. This report adds to our knowledge of chitinozoan paleobiogeography and to other recent studies of Middle De- vonian palynology in northern Iberia.


Introduction
While the chitinozoan record of the Silurian and Early Devonian of northern Spain is relatively well-known, with over one hundred taxa described from the area (Cramer, 1964(Cramer, , 1966a(Cramer, , 1966bCramer and Díez, 1978;Díez and Cramer, 1978;Richardson et al., 2001), the Middle Devonian strata has received next to no formal study. Only one paper, concerned more with land plant spores, has been written, which figured, but did not describe, three specimens of "Ancyrochitina sp." (Cramer, 1969). See Section 4 for the present authors' opinion of the attribution of these specimens. This is not for lack of suitable deposits or geological information; the lithostratigraphy has received extensive study, and units are often well exposed and present various features of interest. These deposits consist of a clastic unit sandwiched between thick carbonate deposits, representing a marine onshore-offshore transect, and are very well described sedimentologically, while they are also well age-constrained by conodont and land plant spore biostratigraphy. The sequence is known to have been deposited around a group of small islands isolated from the supercontinents of Laurussia and Gondwana (Torsvik and Cocks, 2013) and also spans the Kačák Event (García-Alcalde, 1998;Askew and Wellman, 2018), a major global extinction event with no definite cause. The present study aims to comprehensively describe the chitinozoan taxa from this sequence for the first time, comment on the assemblage's biogeographical similarities/ differences with contemporary assemblages and add to other recent studies concerning the palynology of this sequence.

Geological setting
The present study utilizes the same samples as Askew and Wellman (2018) and the geological setting of the material is described in detail therein. In short, the Naranco, Huergas and Gustalapiedra formations, the subjects of the present study, are laterally equivalent and are found in the Asturias, León and Palencia provinces, respectively. Limestone formations occur stratigraphically below (the Moniello, Santa Lucía and Polentinos formations) and above (the Candás, Portilla and Cardaño formations) and again are all lateral equivalents (Fig. 1). The Naranco, Huergas and Gustalapiedra formations represent a disruption of carbonate deposition, replacing it with clastic sediment. The formations comprise a coarse, sandstone lower unit and a more mixed upper unit with interspersed beds of sandstone and siltstone (with occasional limestone beds), indicating that clastic input was reduced to some extent as compared with the lower unit. Detailed descriptions and environmental interpretations of the variety of facies present can be found in García-Ramos (1978),w h i l e Gibbons and Moreno (2002) provide a useful overview of the geology of this region.
Remains of the marine macrofauna of the Naranco, Huergas and Gustalapiedra indicate an Eifelian-Givetian age , which can be refined by conodont faunas recovered from the surrounding limestones. These indicate an age range for the formation from the Polygnathus costatus costatus zone (middle Eifelian; ca. 391-392 Ma) to the lower Polygnathus varcus zone (early, but not earliest, Givetian; ca. 386-387 Ma) (Becker et al., 2012;. The dispersed spore assemblage described in Askew and Wellman (2018) supports this, though suggests the palynomorph-bearing deposits were deposited within a short interval of a little over 1.5 million years (ca. 386-387.7 Ma), entirely within the early Givetian lemurata-langii Assemblage Zone (Geminospora lemurata Interval Zone) (Breuer and Steemans, 2013). See Fig. 2 for the locations of this study's sample sites within northern Spain. Details of the 30 localities used and samples taken for this study are given in Askew and Wellman (2018).

Material and methods
Two periods of field work were undertaken, in which 130 rock samples were collected, covering the outcrop area of the Naranco, Huergas and Gustalapiedra formations. Dark, fine-grained beds of siltstone and mudstone, interspersed within the layers of sandstone that dominate  the formations, were the focus of palynological sampling. Sample sites were mostly either single exposures or short sections which yielded a few consecutive samples. Three relatively long sections, measuring approximately 35, 70 and 330 m, were also analyzed, with samples taken at intervals throughout the exposure wherever good mud-or siltstone beds were seen. Sample details are given in Askew and Wellman (2018). Samples were processed using standard HCl-HF-HCl acid maceration techniques (Slater et al., 2015;Wellman, 2018a), followed by using a 15 μm sieve to remove some fine mineral residues. Heavy liquid centrifugation using ZnCl 2 removed any remaining mineral matter. Assessment of the resulting kerogen samples identified three barren samples and one very poorly preserved one; these were not processed further.
Preservation is generally good but thermal maturity is variable, with a TAI value between 2 and 4 (Traverse, 2008). Oxidization of all the samples was undertaken with Schulze's solution for 5-60 min, except four very thermally mature samples that were oxidized for around 19 h. Quantitative analysis was undertaken by spiking the samples with tablets containing modern Lycopodium spores at a ratio of one tablet per millilitre of organic residue. An equation derived from that given by Stockmarr (1971) then gives the number of palynomorphs per gram of processed rock, subject to the standard deviation given for the tablet batch (all tablets used here came from the same batch). Between two and four strew mounts were made for each sample, with 113 samples being counted using a standard 200 count. Once a total of 200 palynomorphs was reached, additional species were noted as rare. In cases where the slide did not contain 200 palynomorphs, every palynomorph was counted. Such slides are evident in the count data found in Askew (2019). Slides were examined using a light microscope and specimens photographed using a Meiji Techno Infinity 1-5C camera attached to a Meiji Techno MT5300H transmitted light microscope.
SEM materials were prepared in two ways; (i) picking individual chitinozoans for attachment to SEM stubs using carbon tabs and (ii) strew mounting. This involved fixing a glass cover slip to an SEM stub using Araldite® brand epoxy resin, before a palynomorph sample was strewn over it and dried. All stubs were gold-coated using an Edwards S150B sputter coater. All imaging was carried out using a Philips XL-20 SEM.

Systematic paleontology
The chitinozoans described below are arranged alphabetically by genera and treated as genera and species in accordance with the ICZN (1999), though the true origin of chitinozoans is not certain. Species with an uncertain identification are signified cf. or?, after Matthews (1973). A limited synonymy is provided for species recorded previously from the Naranco, Huergas and Gustalapiedra formations. This amounts to three specimens of Ancyrochitina sp.
figured by Cramer (1969). These figured specimens are here placed with species found in the present study. Dimensions are given for each species and follow the scheme of Paris (1981) where L = total length; Dp = chamber diameter; Lp = chamber length; Dc = neck diameter; Ln = neck length. Dimensions are given in micrometres (μm) and no correction factors allowing for compression are used. Where three numbers are given for a dimension these correspond to the minimum value (arithmetic mean) and maximum values. The number of specimens measured is given. Occasionally not all measurements could be obtained from all specimens owing to damage, though in all cases specimens retained sufficient diagnostic features to facilitate identification. Where this is the case, the number measured corresponds to the highest number of measurements for any one dimension (e.g. up to 5 specimens measured, where some dimensions will have five measurements, others less than five owing to loss of features, such as neck flanges, in some specimens). Species occurrences are given as locality code characters for brevity (see Askew and Wellman (2018) for explanation). Previous records are taken from consultation of the John Williams Index of Paleopalynology (for details see Riding et al. (2012)) and are summarized for taxa with worldwide or near-worldwide distributions. Materials (rock samples, residues and slides) are housed in the collections of the Centre for Palynology, Department of Animal and Plant Sciences, University of Sheffield, UK. Figured specimens are located using an England Finder.
Genus Alpenachitina Dunn and Miller, 1964 Type species: Alpenachitina eisenacki Dunn and Miller, 1964 Alpenachitina eisenacki Dunn and Miller, 1964 Plate I,1 Description: Chamber ovoid to cylindrical with a flat to slightly convex base. The neck occupies~20% of the total length. The flexure is conspicuous and the aperture flares, gaining~40% extra width at the aperture versus its narrowest point. Granulate ornament seen near aperture. Flanks bear large processes with up to fourth order dichotomous branching arranged in rings on the margin and shoulder, up to 4 μm wide and up to 20 μmlong.
Remarks: Alp. eisenacki as originally described has a proportionately longer, non-flaring neck bearing an additional ring of processes. If the synonymy of this species with Alp. ontariensis Legault, 1973, as proposed by Almeida- Burjack and Paris (1989),isacceptedthentheflaring neck without true oral processes of the specimens seen here falls within the diagnosis of Alp. eisenacki, despite the poor preservation of the present material. This is despite the generic diagnosis of Alpenachitina describing three distinct "crowns" of processes on the margin, shoulder and neck.
Occurrence: Sites 3, 16. Previous records: Considered characteristic of the Middle Devonian, though with a much larger reported range. Known from late Pragian to Frasnian strata and with an almost worldwide distribution, being absent from Europe, Asia and Australia.
Genus: Ancyrochitina Eisenack, 1955a Type species: Ancyrochitina ancyrea (Eisenack, 1931) Ancyrochitina ancyrea (Eisenack, 1931)? Plate I,2 Description: Chamber conical with a flat to slightly convex base. The neck may occupy~50% of the total length, though damage makes this unclear. The flexure is conspicuous and the aperture flares, though damage precludes measurement of its magnitude. Possible microgranulate ornament observed on surface. Processes seen on margin with up to fifth order dichotomous branching, up to 24 μmlong.
Remarks: The specimens recovered here are damaged, precluding a more confident identification. Those characteristics which are preserved, namely the small surface ornament and ramified processes, match previously described specimens of Anc. ancyrea, a well-known species.
Occurrence: Sites 14, G. Previous records: Anc. ancyrea is reported from Middle Ordovician to middle Frasnian strata and with an almost worldwide distribution, being absent from Australia.
Ancyrochitina cf. flexuosa Burjack, 1996 Plate I,3 Description: Chamber lenticular to conical with a flat to slightly convex base. The neck occupies~46% of the total length. The flexure is conspicuous and the aperture may flare, gaining~17% extra width at the aperture versus its narrowest point when it does so. Processes seen with up to fourth order dichotomous branching arranged in a ring on the basal margin and irregularly arranged on the neck towards the aperture, up to 8 μm wide and up to 46 μm long, smaller on the neck. Remarks: Anc. flexuosa as originally described has a flaring neck and a discrete ring of oral processes, in contrast to the occasional flare and irregular arrangement of the neck processes seen in these specimens, hence their uncertain identification. The neck processes on the specimens seen here are rather more disorganized. The shape of these specimens is somewhat variable, particularly the neck, however this was also seen by Burjack (1996) and allowed for in the original description. Many of the specimens seen here resemble Anc.

Ancyrochitina taouratinensis
Description: Chamber conical to slightly lenticular with a flat to slightly convex base. The neck occupies~37% of the total length. The flexure is conspicuous and the aperture flares, gaining~27% extra width at the aperture versus its narrowest point. Granulate and small spinose ornament observed on surface. Processes seen on margin with up to fourth order dichotomous branching, 3-8 μm wide and 7-23 μm long. Remarks: Specimens included in this taxon exhibit various degrees of damage resulting in loss of processes, though some remnant remains in all cases. The species is distinguished from Anc. ancyrea by its granulate surface ornament.
Ancyrochitina cf. tomentosa Taugourdeau and de Jekhowsky, 1960 Plate I,13 Description: Chamber conical with a flat to slightly convex base. The neck occupies~40% of the total length. The flexure is conspicuous and the aperture does not flare. Microgranulate and spinose or elongated filamentous ornament seen on neck up to 6 μm high. Processes seen on margin with dichotomous branching, up to 11 μmlong.
Remarks:Thethin,filamentous neck ornament of this species serves to distinguish it from other members of the genus. The specimens seen here are damaged, as seen in the figured specimen, rendering their identification uncertain. Characteristics described here are not necessarily seen on all specimens owing to damage, though their distinctive neck ornament is sufficient to refer them to this taxon. Their squat shape resembles Anc. t. compactus Taugourdeau, 1965. Occurrence: Sites 3, 7. Previous records: Anc. tomentosa is reported from middle Llandovery to Frasnian strata and with an almost worldwide distribution, being absent from Australia.

Genus: Angochitina Eisenack, 1931
Type species: Angochitina echinata Eisenack, 1931 Angochitina capillata Eisenack, 1937 Plate I,5 Description: Chamber spherical to ovoid in shape. The neck occupies 39% of the total length. The flexure is conspicuous and the aperture may flare, gaining~15% extra width at the aperture versus its narrowest point when it does so. Microgranulate, granulate and spinose ornament up to 5 μm high evenly distributed across surface. Remarks: The species was originally described as having an ovoid chamber, though Jenkins (1969) describes and figures specimens with a more variable chamber shape, including more spherical ones very similar to those seen here. Some previously recorded specimens appear to have thicker spines than those seen here, thought his measurement is never specified in the original description.
Occurrence: Sites 3, 10, G. Previous records: Reported from Middle Ordovician to Famennian strata and with a worldwide distribution. Description: Chamber spherical to ovoid in shape. The neck occupies 40% of the total length. The flexure is conspicuous and the aperture flares, gaining~20% extra width at the aperture versus its narrowest point. Spinose ornament seen on surface with rare branching spines and larger branching processes, up to 6 μmw i d ea n du pt o2 2μm long. Simple ornament more common on the neck, branching spines and processes are largely limited to the chamber. Remarks: These specimens seem to have a proportionately shorter neck than the species as originally described, however they are well distinguished from Ang. milanensis by possessing fewer wide-based spines and granulae. This is seen in both the transmitted light and SEM images given in the plates; no more than three wide-based structures are seen in Ang. devonica specimens here, while they are common in Ang.
Previous records: Reported from Middle Ordovician to Famennian strata and with an almost worldwide distribution, being absent from Australia. Collinson and Scott, 1958 Plate I,7;Plate III,2-3

Angochitina milanensis
Description: Chamber ovoid. The neck occupies~33% of the total length. The flexure is conspicuous and the aperture often flares, gaining 28% extra width at the aperture versus its narrowest point when it does so. Ornament of spines, simple or with up to third order dichotomous branching and up to 27 μm long, with wide-based, branching granulae mostly limited to the chamber, up to 23 μm wide and up to 15 μmlong. Remarks: This species is distinguished in this assemblage by its wide-based elements. Ang. devonica can exhibit the same ornament type but it is usually smaller and the overall shape is more elongate.

Dimensions
Occurrence: Sites 3, 7, 13, G. Previous records: Reported from Early Silurian to early Famennian strata and with an almost worldwide distribution, being absent from Asia.

Angochitina sp. A Plate I,14
Description: Chamber ovoid. The neck occupies~30% of the total length. The flexure is conspicuous and the aperture flares, gaining 30% extra width at the aperture versus its narrowest point. Microgranulate ornament seen on the chamber and neck, with a spinose ornament clustered on the shoulder up to 5 μmhigh.
Remarks: These specimens most closely resemble Ang. capillata in this assemblage, except for their more ovoid chamber and particularly the restricted distribution of these specimens' spinose ornament. No existing species could be found to which these specimens could be confidently assigned.
Occurrence: Sites 3, 7. Angochitina sp. B Plate I,12 Description: Chamber ovoid. The neck occupies~30% of the total length. The flexure is conspicuous and the aperture flares, gaining 40% extra width at the aperture versus its narrowest point.

Remarks:
The distinctive ornament of large cone-like elements distinguishes these specimens, for which no analog could be found in the literature. The ornamental elements could represent the bases of eroded spines but their rounded tips and lack of obvious signs of breakage do not support this interpretation.
Occurrence: Sites 13, G. Angochitina sp. C Plate I,8 Description: Chamber very slightly ovoid. The neck is unknown. Sparse granulate and spinose ornament up to 4 μm high, distributed across surface. A single multi-rooted spine is seen on the shoulder.
Remarks: This single specimen is unique in this assemblage in possessing multi-rooted spine ornamentation. The specimen strongly resembles Ang. communisJenkins, 1967, though it is too damaged to permit a confident identification.
Occurrence:Site10. Angochitina sp. D Plate I,10;Plate III,4 Description: Chamber ovoid. The neck occupies~30% of the total length. The flexure is conspicuous and the aperture flares, gaining 10% extra width at the aperture versus its narrowest point. Granulate and spinose ornament seen, dense on neck and occasionally seen on chamber, up to 4 μm high. Those seen on the chamber may be broken spine bases. Remarks: The specimens found here resemble Ang. toyetae Cramer, 1964 figured by Legault (1973) as pl. VI, fig. 10, though that specimen is acknowledged to be damaged. The same is probably true of the specimens seen here.

Dimensions:L=69
Remarks: These specimens present few diagnostic features, other than their unusual proportions. The overall shape of these specimens is very similar to the genus Desmochitina Eisenack, 1931, but they do not possess the correct surface texture.
Occurrence: Sites 3, 10, G. Genus: Fungochitina Taugourdeau, 1966 Type species: Fungochitina fungiformis (Eisenack, 1931) Fungochitina cf. lata (Taugourdeau and de Jekhowsky, 1960) Plate II,1 Description: Chamber conical with a more or less rounded margin and a flat to convex base. The neck occupies~35% of the total length. The flexure is conspicuous and the aperture flares, gaining~20% extra width at the aperture versus its narrowest point. Fine microgranulate ornament seen across surface with occasional larger elements up to 2
Occurrence: Sites 3, 4, 598, G. Previous records: Reported from Late Silurian to late Frasnian strata and with an almost worldwide distribution, being absent from Asia and Australia.
Fungochitina cf. pistilliformis (Eisenack, 1931) Plate II,3 Description: Chamber conical with a flat to slightly convex base. The neck occupies~30% of the total length. The flexure is conspicuous and the aperture does not flare. Microgranulate, granulate and spinose or elongated filamentousornamentseenonneckupto5μmhigh. Remarks: Conochitina pistilliformis as described by Eisenack (1931) has a proportionately much longer neck than the specimens seen here, though many of these specimens are damaged and missing part of the neck. They are distinguished from F. cf. lata primarily by their restricted covering of ornamental elements, seen only on the neck.
Remarks: This species is often extremely numerous in this material, possibly because of the ease with which damaged specimens can be identified by the prominent operculum. This may also be the reason for its seemingly large stratigraphical range elsewhere (see below).
Occurrence: Sites 3, 7, 120, 599, P. Previous records: H. glabra is reported from Late Ordovician to Late Permian strata and with an almost worldwide distribution, being absent from Asia, though it is often considered characteristic of the Late Devonian (Paris et al., 2000). Reports of the species occurring more widely may be instances of homeomorphs. Remarks: This single specimen is placed in the genus Lagenochitina by its glabrous surface and ovoid chamber. The lack of SEM imagery of the specimen might render this identification doubtful, but no trace of ornament could be seen even under 100 × magnification. No specific identification has been attempted owing to the specimen's damaged state.
Occurrence:Site10. Genus: Ramochitina Sommer and van Boekel, 1964 emend. Paris et al., 1999 Type species: Ramochitina ramosi Sommer and van Boekel, 1964 Ramochitina corniculata (Laufeld, 1974)? Plate II,8-9 Description: Chamber lenticular with a more or less convex base. The neck occupies~35% of the total length. The flexure is conspicuous and the aperture may flare, gaining~10% extra width at the aperture versus its narrowest point when it does so. Broken process bases seen on chamber, roughly arranged in longitudinal rows, up to 5 μm wide and Remarks: These specimens are doubtfully assigned to the species on account of their relatively short necks and few spines, both attributed to damage. They do, however, bear a great similarity to the holotype of R. corniculata depicted in fig. 47B of Laufeld (1974) and the specimen figured here in Pl. II, 8 shows one process curved orally, characteristic of the species.
Occurrence: Sites 3, 7. Previous records: R. corniculata is reported from the late Llandovery to middle Wenlock of Belgium (Verniers and Rickards, 1979;Verniers, 1999;Verniers, 1981;Verniers et al., 2002)a n dt h eW e n l o c ko f Sweden (Laufeld, 1974;Jenkins and Legault, 1979) and Wales (Verniers, 1999). Grahn and de Melo, 2002? Plate II,10 Description: Chamber a strongly rounded conical shape with a flat base. The neck occupies~50% of the total length. The flexure is conspicuous and the aperture flares, gaining~35% extra width at the aperture versus its narrowest point, though the neck is damaged. Spinose ornament observed arranged in rough longitudinal rows, up to 8 μm high and limited to the aboral end of the chamber around the margin.
Remarks: This single specimen differs from the species as originally described by its lack of spines on the neck and fewer spines generally, though the latter may be due to damage. The general form and proportions of this specimen are comparable with R. derbyi.
Occurrence:SiteG. Previous records: R. derbyi is reported from the early Frasnian of Brazil Grahn and de Melo, 2002).
Ramochitina cf. magnifica Lange, 1967 Plate II,15;Plate III,5 ?1969aAncyrochitina sp.;Cramer, pl. IV, fig. 45 Description: Chamber ovoid. The neck occupies~35% of the total length. The flexure is conspicuous and the aperture does not flare. Ornament of processes with up to third order dichotomous and trichotomous branching, occasionally with wide bases. Processes 1-9 μmwide and 13-36 μm long. Processes arranged in longitudinal crests. Remarks: The specimens found here are considerably smaller than the species as originally described but conform with it morphologically.

Ramochitina sp. A Plate II,5
Description: Chamber conical with a slightly convex base. The neck occupies~20% of the total length. The flexure is conspicuous and the aperture does not flare. Microgranulate and longitudinal crest-like ornament seen on chamber, possibly made up of coalesced smaller elements. One small branching process also seen on margin.
Remarks: This single specimen is reminiscent of R. spinosa (Eisenack, 1932), though it has a shorter neck and smaller ornament, possibly due to damage. Paris et al., 1999 Type species: Saharochitina jaglini (Oulebsir and Paris, 1993) Saharochitina sp. A Oulebsir and Paris, 1993 Plate III.

Plate II,11
Description: Chamber conical with a very slightly convex base. The neck occupies~30% of the total length. The flexure is conspicuous and the aperture does not flare. Surface glabrous.
Remarks: This single specimen strongly resembles Fungochitina?sp. AofOulebsir and Paris (1993), though with a slightly shorter neck. It is this short neck that distinguishes the form from F.? jagliniOulebsir and Paris, 1993, since reassigned as the type species of Saharochitina by Paris et al. (1999) owing to its glabrous surface.
Occurrence:SiteP. Previous records: Oulebsir and Paris (1993) report their specimens from the Early and Middle Ordovician of Algeria.
Genus: Sphaerochitina (Eisenack, 1955a) emend. Paris et al., 1999 Type species: Sphaerochitina sphaerocephala (Eisenack, 1932) Sphaerochitina cf. compactilis Jenkins, 1969 Plate II,12 Description: Chamber ovoid to almost spherical. The neck occupies 20% of the total length. The flexure is conspicuous and the aperture may flare slightly. Surface glabrous. Remarks: Sph. compactilis as originally described has a conical to spherical chamber, however the figured specimens are better described as ovoid in the opinion of the present authors. The species is also described as having a neck occupying up to one-third of the total length, something that is not obvious in the figured specimens. Indeed, the specimens seen here bear a marked resemblance to the specimen figured in pl. 9, fig. 16 of Jenkins (1969).

Remarks:
The specimens seen here do not show the tendency for their ornament to cluster that Taugourdeau (1962) described, though they match the type material in all other particulars including shape, proportions and ornament type.
Occurrence: Sites 3, 4, 7, 10, 598, G. Previous records: Reported from the Early Silurian of Saudi Arabia (McClure, 1988), the Late Silurian of Tunisia (Grignani, 1967), the Frasnian (Taugourdeau, 1962) of Algeria, the Early to Middle Devonian of Brazil (van Boekel, 1966, 1968a, 1968bda Costa, 1971), the early Emsian to early Eifelian of Spain , the Frasnian of France (Moreau-Benoit, 1965) and the Late Devonian of Morocco (Grignani and Mantovani, 1964). Díez and Cramer, 1978 Plate II,16;Plate III,7 Description: Chamber spherical to lenticular in shape. The neck oc-cupies~60% of the total length. The flexure is conspicuous and the aperture does not flare. A sparse microgranulate ornament is seen on the surface, with elements well separated from each other.
Remarks: This single specimen strongly resembles Sph. ricardi in proportions and ornament, except it does not have the conical shape specified for the species. This is interpreted here as due to later distortion disguising the conical vesicle of this specimen, as it does resemble some of the specimens figured by Díez and Cramer (1978). The SEM image seen in Pl. III, 7 shows a specimen with an apparent ornament, however this is believed to be mineral growth.
Occurrence:Site10. Previous records: Reported from the Pragian of Australia (Winchester-Seeto and Carey, 2000) and the early Emsian to early Eifelian of Spain .
Sphaerochitina sphaerocephala (Eisenack, 1932)s e n s uDíez and  Plate II,17-18; Plate III,8 Description: Chamber ovoid. The neck occupies~35% of the total length. The flexure is conspicuous and the aperture may flare, gaining 15% extra width at the aperture versus its narrowest point when it does so. A faint microgranulate ornament is seen on the surface, with elements quite well separated from each other. Remarks: One specimen (pl. II, 17) appears to have a single spine on the shoulder, though this may not be original; no other spines or spine bases are seen on this specimen. This species is distinguished from Sph. cuvillieri in this study by its much fainter, more widely spaced microgranulate ornament. Sph. sphaerocephala as originally described has a much longer neck than the specimens seen here. Díez and Cramer (1978) figure shorter, more squat specimens, referencing Eisenack (1968) which included specimens of a similar shape. These show a marked similarity to those seen in the present study, though they may simply represent part of the considerable variation seen in Sph. sphaerocephala, rather than a distinct form. The SEM image seen in Pl. III, 8 shows a misshapen specimen, probably due to internal mineral growth.
Occurrence: Sites 3, 4, 7, 10, 11, 120, G. Previous records: Sph. sphaerocephala is reported from Early Ordovician to Famennian strata and with an almost worldwide distribution, being absent from Australia. Díez and Cramer (1978) describe specimens similar to those found here from the early Emsian to early Eifelian of Spain and reference specimens found from the Ordovician to Silurian of the Baltic (Eisenack, 1968).

Other forms
Chitinozoan type A Plate II,14 Description: Chamber possibly claviform. The flexure is inconspicuous and the aperture flares, gaining~20% extra width at the aperture versus its narrowest point. Microgranulate and fine spinose ornament up to 6 μm high seen on surface.
Remarks: Clavachitina Taugourdeau, 1966 is described by Paris et al. (1999) as being glabrous, precluding the spinose ornament of this species. This form may belong to Belonechitina Jansonius, 1964 but preservational deficiencies preclude a confident identification in either case. Occurrence:Site3.

Description of the chitinozoan assemblage
The chitinozoan assemblage consists of 29 taxa, including 20 named species identified confidently or uncertainly, assigned to nine named genera. The remaining nine taxa are left in open nomenclature, having no previously identified analog in the literature. Charts of the quantitative abundances of certain taxa are given as Figs. 3-5,withfull charts available as supplementary information. Count data is available at Askew (2019). The assemblages recovered from each sampled site show no significant differences in their species compositions, therefore the community is considered here as a single regional assemblage.
No clear temporal changes are evident in the assemblage. Inceptions do occur through the sampled sections, but they show no evidence of being the result of assemblage composition changes through time. They do not appear in the same order, for instance, and no lateoccurring species become important components of the assemblage. This may be connected to the short time period in which the formations were deposited (see Section 2), leaving no time for chitinozoan assemblages to change significantly. Out of the total assemblage only Angochitina sp. C, Lagenochitina sp. A, Saharochitina sp. A and Sphaerochitina ricardi are not seen in the sections. These are all extremely rare taxa, only occurring in one sample each in this study. That, combined with their lack of biostratigraphical importance, mean their absence in the stratigraphical sections is not considered significant.
Reworking is not evident in the palynomorph assemblages previously described for the deposits (Askew and Wellman, 2018), and the chitinozoans show no obvious differences in preservation and thermal maturity, suggesting no noticeable reworking has occurred with them either. There is also no obvious, systematic reworking of older forms from the comparatively well-studied La Vid Group (Cramer, 1964;Cramer-Díez et al., 1972;Díez and Cramer, 1978) into the present assemblage.
6. Comparison with other chitinozoan assemblages Chitinozoan regional assemblages that are contemporaneous with the present one are relatively well-known, however many reports contain very few species or are poorly age-constrained. Taxonomy can also be a confounding factor, as species concepts have been changed and taxa re-evaluated. Despite this, 28 assemblages (10 from Laurussia and 18 from Gondwana) containing sufficient taxa for a meaningful comparison with the present assemblage have been found.
Two measures of similarity have been used here. The coefficient of similarity (CS) (sensu Clark and Hatleberg (1983)) is expressed as CS =2|x|/(|a + b|) where a and b are the total number of species in assemblages a and b, respectively, and x is the number of species they share. Results below 0.2 indicate low similarity, while results between 0.2 and 0.55 are considered moderate to high (Clark and Hatleberg, 1983). Secondly, the Jaccard Index (JI) can be expressed as JI = | x |/ Fig. 3. Quantitative abundances of selected chitinozoan taxa at the Playa del Tranqueru site. Abundances expressed as numbers per gram of rock. Each horizontal tick represents 10 specimens per gram. The section's lithology is also shown. The "Sphaerochitina sphaerocephala" indicated here represents Sph. sphaercephala sensu Díez and Cramer, 1978 (see Section 4).
(|a|+|b|-|x|) where a and b are the total number of species in assemblages a and b, respectively, and x is the number of species they share. Completely dissimilar assemblages score 0 while identical ones score 1. These measures have been used previously in Paleozoic palynological applications (e.g. Le Hérissé et al., 1997;Wellman, 2018b;Wellman et al., 2013). In all cases, taxa with only a generic assignment and those designated "?" have been excluded from the calculations, while those designated "cf." have been treated as valid identifications where appropriate following individual assessment. All were assessed using modern taxonomic assignments.
Values for these metrics are given in Table 1.Thisassemblageisseen to not be particularly similar to contemporary ones from other parts of the world. It is most similar to that of Anan-Yorke (1974) from Ghana (CS: 0.480; JI:0.316), followed by those of Lange (1967) from Brazil (CS:0.370; JI:0.227) and Urban (1972) from the USA (CS:0.345; JI:0.208), showing no particular geographical bias. These three are fairly high similarity results, but most are moderate or low. At the continental scale, this assemblage is slightly more similar to Gondwana (CS:0.246; JI:0.140) than to Laurussia (CS:0.228; JI:0.129) but the difference is very small, and both continents show only a low to moderate degree of similarity. spinosa and Ramochitina ramosi. The absence of species belonging to Urochitina and Eisenackitina in this analysis is also notable, as both genera were common at the time (Paris et al., 1999). Indeed, only one species of desmochitinid, Hoegisphaera cf. glabra, was recovered here. This form is long ranging and possibly includes a number of homeomorphs.
Laurussia and Gondwana do not have particularly similar assemblages, with few species common on both continents. Despite this, the Spanish assemblage is not very similar to either of them. This may be connected to the isolated position of Spain during the Middle Devonian (see Section 1).

Biostratigraphical implications
Chitinozoan biostratigraphical schemes for the Middle Devonian are somewhat lacking owing to a paucity of comprehensive study of the period (Paris et al., 2000). Assemblages may differ significantly between regions (Miller, 1996), meaning regional schemes should be used with caution here, especially since no previous study of Spain has been performed and it has not been considered in previous schemes.
The formations in which this chitinozoan assemblage was found are well age-constrained by conodonts and dispersed spores (see Section 2) to the middle Eifelian to early Givetian stages, with the palynomorphbearing material occurring in the early Givetian only. The depositional period of the formations is divided into three biozones in the global scheme of Paris et al. (2000).  the upper Middle Polygnathus varcus conodont zone (Paris et al., 2000), however this is younger than the top of the formation is known to be usingconodontdating(seeSection 2). The absence of other proxies prevents a trustworthy biostratigraphical conclusion.
The species identified here, both confidently and tentatively, are generally Middle Devonian species, though certain taxa are not previously recorded from this time period. Ramochitina corniculata is only known from the Silurian, while R. derbyi is known from the Frasnian, but both these species are only tentatively identified here owing to few, damaged specimens and may represent homeomorphs.
Saharochitina sp. A Oulebsir and Paris, 1993 is recorded only from the Ordovician, though it is a very simple form with few distinguishing features and, again, may be a homeomorph. Sphaerochitina compactilis is another simple form, this time known from the Ordovician, with something of a disconnect between its description and the original figured specimens, also only identified tentatively here and possibly another homeomorph. Sph. ricardi is only reported up to the early Eifelian, though that report is from Spain. It is possible the present study represents a genuine expansion in this species' known temporal range. The specificf or mo fSph. sphaerocephala found here, Sph. sphaerocephala sensu Díez and Cramer (1978), is also only known from the early Eifelian of Spain, though the species sensu stricto ranges from the Ordovician to Famennian.
A somewhat varied biostratigraphical picture is presented by this data. Some "out of place" species may be homeomorphs or they may be valid extensions to the ranges of some species. The lack of previous study of this material lends support to the latter conclusion, but the limitations of primarily light microscopy studies such as this one largely preclude a definite say in this matter. None of these potentially discordant taxa are known to be particularly stratigraphically important in any case. Reworking is unlikely to account for these taxa; it is not evident in the rest of the assemblage and it cannot account for the supposedly Fig. 5. Quantitative abundances of selected chitinozoan taxa at the Crémenes-Las Salas site. Abundances expressed as numbers per gram of rock. Each horizontal tick represents 10 specimens per gram. The section's lithology is also shown. The "Sphaerochitina sphaerocephala" indicated here represents Sph. sphaercephala sensu Díez and Cramer, 1978 (see Section 4). younger species recorded here. These unexpected taxa add to the unusual chitinozoan assemblage recovered here from Middle Devonian Spain.

Conclusions
• A diverse assemblage of mostly well-preserved chitinozoans has been recovered from Middle Devonian deposits in northern Spain.
• The assemblage does not show strong similarity to contemporary assemblages from either Laurussia or Gondwana, reflecting the isolated position of Iberia during the Middle Devonian. Various taxa also appear to have occurred earlier in Spain than elsewhere, with others persisting for longer.
• The chitinozoan assemblage does not refute the Middle Devonian age of the formation determined by conodont and dispersed spore data.

Acknowledgment
This work was financially supported by a NERC (Natural Environment Research Council, UK) studentship awarded to AJA through the ACCE (Adapting to the Challenges of a Changing Environment) Doctoral Training Partnership (Grant no. X/008677-12-1). AJA is grateful to his PhD supervisor Prof. Charles Wellman and to Dr. Stephen Stukins (Natural History Museum, London) for providing access to the JWIP. The authors are grateful to Dr. Florentin Paris for invaluable advice and guidance concerning the taxonomic assignments used here, and to Dr. Anthony Butcher and one anonymous reviewer for reviewing earlier versions of this manuscript.