Specimen: NHMUK PV R 9731 * axis and four post-axial cervical vertebrae. Two incomplete axial ribs* the left fourth cervical rib* and two cervical rib fragments.

Locality: Charmouth * Dorset * England * United Kingdom.

Horizon: Charmouth Mudstone Formation* Lias Group.

Age: Sinemurian or early Pliensbachian.

Description: All of the preserved vertebrae (NHMUK PV R 9731; Figs 6–8) come from the cranial-most region of the neck of a juvenile thalattosuchian (due to the parapophyses being borne low on the centrum* Andrews 1913). A thalattosuchian identification can be hypothesized due to its characteristic hourglass shape (e.g. Fraas 1902 * Arthaber 1906 * Andrews 1913 * von Huene 1925 * Westphal 1962 * Herrera et al. 2013c * Martin et al. 2019a; Figs 6 * 7) and the lack of any evidence of hypapophyseal laminar processes or ‘knobs’ at the cranial half of the centra ventral surfaces that correspond to the hypapohysis (such as in goniopholidids and pholidosaurids; Hua et al. 2007 * Martin et al. 2016b; Fig. 6)* while the presence of dichocephalous axial ribs has only been reported for metriorhynchoids amongst thalattosuchians (e.g. Andrews 1913 * Westphal 1962 * Johnson et al. 2020a * Young et al. 2024; Fig. 8). All vertebrae are small in size and have open neurocentral sutures (Figs 6 * 7). The specimen was donated to the NHMUK in 1967 by Mr. Barney Hansford. The specimen is stated to have been collected from Charmouth* Dorset * but there is no more locality information. As such* we cannot be sure what member of the Charmouth Mudstone Formation it originated from.

The first preserved cervical vertebra is the axis (C2). The centrum is 15 mm long (although the cranial end of the centrum is incomplete)* with the articular faces being 12.6 mm tall and 10 mm wide. The total height of the vertebra (including the neural arch) is 33 mm. The main body of the centrum is very strongly compressed mediolaterally* with numerous neurovascular foramina being visible in lateral view (Fig. 6A). Only the caudal articular surface is preserved* with it being amphicoelous (noticeable concavity is present). There is a well-defined hypapohyseal ridge that extends ventrally below the articular surfaces (Fig. 6C). This ridge is not blade-like (i.e. not laminar)* nor does it have the ‘knob’-like appearance seen in pholidosaurids and goniopholidids (e.g. Hua et al. 2007 * Martin et al. 2016b). Thalattosuchian hypapophyses are typically poorly developed* if they are present at all* being mediolaterally thick (i.e. not forming a laminar blade) and following the contour on the centrum (see the figures in: Andrews 1913 * Herrera et al. 2013c * Young et al. 2013b * Parrilla-Bel and Canudo 2015* Martin et al. 2019a * Sachs et al. 2019 * 2020 * 2021 * Wilberg et al. 2023). Here* the hypapophyseal keel forms a largely horizontal shape in lateral view* on both the axis and the C3 (Fig. 6A). Only the left parapophysis is preserved* and then only partially. It is borne very low on the centrum* at the cranial ventrolateral edge.

The axis neural arch is largely complete* although the preservation is better on the left-side than the right (Fig. 7). The prezygapophyses are large* broad* and inclined craniomedially* overhanging the cranial surface of the centrum (Figs 6 * 7). The diapophyses are situated slightly caudal to the prezygapophyses and are short and directed predominately ventrally. The postzygapophyses are situated dorsal to the neural canal* and are large* broad* and inclined ventrolaterally. On the left-side of the neural arch a cranial process is preserved dorsal to the prezygapophyses* on the right-side the preservation is such that it is not present. This cranial process has an articulation surface on its dorsal margin* continuing on to the neural arch. We posit that this surface is for the reception of the axis neural spine* much like in the ‘ Steneosaurus ’ hulkei Andrews* 1913 holotype (NHMUK PV R 2074). Such a cranial process is not common amongst thalattosuchians* especially metriorhynchoids (e.g. Fraas 1902 * Arthaber 1906 * Andrews 1913 * Sachs et al. 2020).

The post-axial cervical vertebrae are all amphicoelous* with deep concavities on their articular surfaces (Figs 6 * 7). None preserve complete neural spines* with differences in breakage between them. The right-hand side of the vertebrae have the poorest preservation* just like the axis (Fig. 7). The second preserved cervical vertebra is the first post-axial (C3)* which has a centrum 15.7 mm long* with the articular faces being 14 mm tall and 11.9 mm wide. The total height of the vertebra (including the neural arch) is 41 mm (although the tip of the neural spine is broken). The third preserved cervical vertebra is the C4* which has a centrum 13.5 mm long (but the caudal end of the centrum is incomplete)* with the articular faces being 15 mm tall and 11.9 mm wide. The total preserved height of the vertebra is 31.3 mm tall* although the neural spine is incomplete* and the centrum is distorted. The fourth preserved cervical vertebra is the C5* which has a centrum 16.5 mm long* with the articular faces being 13.7 mm tall and 13.3 mm wide. The total preserved height of the vertebra is 38 mm tall* although the neural spine is incomplete. The fifth preserved cervical vertebra is the C6* which has a centrum approximately 18 mm long* with the articular faces being 15 mm tall and 13.8 mm wide. The total preserved height of the vertebra is approximately 31 mm tall* although the neural spine is incomplete.

As with the axis* the hypapophysis of the C3 is well defined* extending ventral to the articular surfaces but is not blade-like.

The hypapophysis on the C4 is lower than the C2 and C3 vertebrae* with the hypapophyses on the C5 and C6 becoming progressively less prominent. The large parapophyses* borne low on the centra* stop the vertebra from being considered ‘hourglassshaped’. Moreover* the compression of the centra becomes less pronounced from the C4 to the C6* again contributing to the lack of an hourglass morphology. The diapopohyses are short and ventrally inclined for the C3–C5 (like the axis); however* the C6 has noticeably longer diapophyseal processes. The prezygapophyses of the vertebrae are all well developed* oriented dorsomedially* are located dorsal to the neural canal and project cranially beyond the centrum. The postzygapophyses are also well developed* oriented ventromedially* are located dorsal to the neural canal and project caudally beyond the centrum.

The C4 vertebra has the left rib glued to the parapophysis and diapophysis (Fig. 7C * H). There are four other cervical rib fragments* two belong to the axial ribs and two are postaxial cervical ribs (Fig. 8). All of the ribs are dichocephalous (‘double-headed’) with well-defined capitular and tubercular processes. The morphology of the axis ribs is consistent with metriorhynchoids (e.g. Andrews 1913: text-fig. 67A)* with the cranial end of the rib having a cranially oriented capitular (parapophyseal) process and a craniodorsally oriented tubercular (diapopohyseal) process. Of the postaxial cervical ribs* only the one glued to the C4 vertebra is well preserved. It has well-developed capitular and tubercular processes that are glued to the parapophysis and diapophysis respectively. The main body of the rib has a flattened morphology* with an incomplete cranial process and a longer caudal process.

Given that the specimen is from a small juvenile* it is hard to give a more exact identification than Thalattosuchia indeterminabilis. Given the presence of dichocephalous axial ribs* we could posit a metriorhynchoid identification (Andrews 1913 * Johnson et al. 2020a * Young et al. 2024). Although it should be noted that the morphology of axial ribs of early diverging thalattosuchians* such as Turnersuchus (Wilberg et al. 2023) * is unknown. In teleosauroids* there is a ‘bump’ on the dorsal margin of their axial ribs (Andrews 1913). This suggests that the tubercular process was reduced in teleosauroids* or possibly unossified. We cannot preclude the possibility that early teleosauroids did indeed have dichocephalous axial ribs. Therefore* we consider this specimen (NHMUK PV R 9731) to be best considered Thalattosuchia indeterminabilis.