Phymosoma granulosum (Goldfuss, 1829)

Schlüter, Nils, Kutscher, Manfred, Smith, Andrew B., Jagt, John W. M. & Lees, Jackie A., 2012, Late Cretaceous phymosomatids and the true identity of Cidarites granulosus Goldfuss, 1829 (Echinoidea, Phymosomatoida), Zootaxa 3271, pp. 17-30 : 18-28

Phymosoma granulosum (Goldfuss, 1829)

( Figs. 1 A–B; 2A–D; 3A–I; 4B–D; 5A; 6A–D, H)

Cidarites granulosus Goldfuss, 1829: 122 , pl. 40, fig. 7.

? Cyphosoma Milleri Bronn in Bronn & Roemer, 1851: 186 (pars), pl. 29, figs. a–c.

Phymosoma granulosum — Lambert & Thiéry 1911: 225.— Ravn 1928: 55 (pars), pl. 6, figs. 6–10 only.

? Phymosoma corollare — Ravn 1928: 57.

Gauthieria cf. pseudoradiata — Kutscher 1973: 112 (pars), figs. 24, 25.

Rachiosoma corollare sensu Cotteau, 1865 —van der Ham & van Birgelen 1992: 49, pl. 2, figs. 1, 2.

Rachiosoma granulosa — Kutscher 1985: 525, pl. 2, figs. 7–12, 2003: 9, pl. 6, figs. 1–5.

non Cyphosoma granulosum — Woodward 1856: 1.—Cotteau 1864: 685, pl. 1169 (= Phymosoma ravni sp. nov.).

non Phymosoma granulosum — Desor 1855: 87.—Wright 1870: 129, pl. 23, fig. 2.—van der Ham et al. 1987: 22, pl. 3, fig. 2.— Smith & Wright 1996: 286, text-fig. 100; pl. 99, figs. 1–6.— Jeffery 1997: 672, text-fig. 5a, b; pl. 3, figs. 1–5.— Smith et al. 1999: 96, pl. 2, figs. 3, 4.— Smith & Jeffery 2000: 86, text-fig. 32D, E.— Jagt 2000: 229, pl. 9, figs. 1–6; pl. 10, figs. 13, 14 (= Phymosoma ravni sp. nov.).

non Phymosoma granulosum — Savchinskaya 1974: 312, pl. 95, figs. 7–10 (= Gauthieria sp.).

non Phymosoma granulosum — Geys 1980: 204, figs. 1/7, 2/1–4 (= Phymosoma ravni sp. nov.).

Types. Two syntypes were available to Goldfuss; one of these is lost. The surviving specimen (RFWUIP 1335; see Fig. 2 A–D) is here designated lectotype; it allegedly came from Maastricht. However, as Jagt (2000: 230) noted, this is doubtful, since the light greyish, fine-grained limestone matrix of the lectotype is of a type which is unknown from the Maastrichtian type area. In order to try and establish the age and provenance of this specimen, a matrix sample was processed for nannofossil content by one of us (JAL). The nannoflora recovered indicates nannofossil Subzone UC 20c BP (base of the subzone = base of Arkhangelskiella maastrichtiensis ; biozonation scheme of Burnett et al. 1998), i.e., an late Maastrichtian age. The possibility that this sample could be Danian (early Paleocene) in age, with reworked Maastrichtian nannofossils, seems unlikely, as there are no exclusively post-Maastrichtian taxa present, nor any post UC 20c BP age-indicative taxa. The provenance of the echinoid test could not be determined on the basis of this matrix sample, although the assemblage likely is typical of northern high- to intermediate latitudes, based on the presence of A. maastrichtiensis and on the absence of members of the end-Cretaceous Micula lineage, which are seen at lower latitudes.

Occurrence. Lower Maastrichtian of Rügen, northeast Germany; lower and upper Maastrichtian of eastern Denmark (Møns Klint, Stevns Klint); lower upper Maastrichtian of the Aachen-Vijlen area, northwest Germany / southeast Netherlands.

Material. GZG.STR.018526, 0 18527, 018531–018536 and 0 18538, all from the lower Maastrichtian of Jasmund, Rügen (leg. M. Kutscher), MB.E 6674 (locality unknown, probably Rügen), MB.E 6866 and 6867 ( Trigonosemus pulchellus Z one, lower Maastrichtian, Promoisel Quarry, Rügen).

Diagnosis. Test large, flattened, with slightly sunken peristome; pentagonal apical opening of medium size; interambulacral plates slightly wider than ambulacral plates, width of adapical interambulacral plates in large-sized tests twice their height, with concomitant rapid decrease in primary tubercle size. Interambulacra in large-sized tests with very well-developed secondary tubercles adorally, as well as adapically; more or less irregularly biserial pore zones occasionally present adapically; short phyllodes present.

Description. Most specimens used for biometric analysis (GZG.STR.018526, 0 18527, 018531–018536 and 018538) are more or less distorted, making measurements of test size and peristomial and periproctal openings approximate. Medium-sized tests (up to 54 mm in diameter) are depressed in profile and circular in outline, test height being c. 44–53 per cent of test diameter. The ambitus is at mid-height or slightly below. The tubercles are crenulate and imperforate.

Peristome. The circular peristome is faintly sunken; its diameter equalling 32–39 per cent of test diameter. Buccal notches are feebly developed. Lantern elements are only fragmentarily preserved in GZG.STR.018534 (a flint nodule, with associated facetted spines).

Apical disc. The apical disc is inferred to have been monocyclic, based on the outline of the pentagonal opening, which is elongate and indents the posterior interambulacrum. The length of the apical opening equals 35–44 per cent of test diameter, the width c. 35 per cent. A monocyclic arrangement can also be inferred by the size and shape of two genital plates preserved in GZG.STR.018538 (see Fig. 6 H). These plates are laterally elongated and pentagonal, the inner margin being smooth and gently concave. The upper surface of the larger plate shows some tubercles and delicate granules. Based on their shape, both would fit into the projecting margins of the apical opening. Gonopores are large, i.e. 1 mm in diameter at plate lengths and heights of 4.5/5.0 and 2.0/2.6 mm, respectively.

Ambulacra. Plate compounding is of phymosomatoid type. The mamelons equal 30–33 per cent of areole diameter. At a test diameter of 47 mm there are 12 or 13 tubercles in a row; in a specimen of c. 47.5 mm in diameter there are 13 to 15. The tubercles occupy most of the plate. The width of the ambulacra is c. 25 per cent of test diameter at the ambitus. The ambulacra taper adapically and, to a somewhat lesser extent, adorally. After the sixth or seventh plate, the tubercles decrease rapidly in size (see Fig. 1 B). At the peristome, short phyllodes are present, which are formed of 7–9 pore pairs at a test diameter of c. 47 mm. On the adapical surface, biserial pore pairs occur only irregularly and only very late (after c. 46–50 pore pairs); either the pore zones are strictly uniserial, or on a few plates, the vertical axes of pore pairs are offset to each other. In the latter case, these pore zones return to a uniserial arrangement closer to the apex. The amount of displacement of pore pairs to each other is highly variable, ranging from only very limited, so that the pore pairs are still in close proximity in a vertical line ( Fig. 6C), to markedly so that the pore pairs are nearly horizontal to each other ( Fig. 6D). This variation is often seen even in different ambulacral zones of a single individual. The pores are arranged in arcs of 5 or 6 pairs at the ambitus. The most adapical 3 to 4 plates consist of an alternation of demiplates and quinquegeminate plates. At the ambitus, the pores of a pair are separated by a prominent ridge. The ambital pore pairs are obliquely oriented; in adapical and adoral direction they are more or less horizontal. At the ambitus, the space between the pore pairs is relatively large and often covered in minute granules. The tubercles are encircled by small bands of miliary granules and small tubercles; in smaller-sized individuals (c. 39 mm test diameter) these granules are missing adradially. The tubercles at the ambitus have radiating grooves, marking plate sutures on their adradial side. The primary tubercles in a column are contiguous ambitally; above and beneath, they are separated by one or two rows of granules. The first 3 to 4 perioral plates have small, yet prominent, tubercles at the adapical corner of the adradial margin, which exceed the size of the granules at the interradial junctions of the ambulacral plates (see Fig. 1 B).

Interambulacra. The interambulacra are about 34 per cent of test diameter at the ambitus. There are 13 plates in a row in individuals of 34 mm in diameter, 13–15 at a size of 47 mm. The tubercles bear medium-sized mamelons that measure 30–36 per cent of the areole diameter. The areoles are oval to circular in outline. The tubercles are confluent at the ambitus and subambitally in smaller individuals (c. 39 mm test diameter); above the ambitus, tubercles are separated by a small zone of minute granules running across the adapical margin of each plate.

In larger specimens (c. 50 mm test diameter), the tubercles are mostly separated by a small zone of granules and the interambulacral plates become rapidly wider than tall above the ambitus. The perradial areas are filled by a moderately broad band of granules, in a zigzag line. The first 4 to 5 plates on the oral side bear 2 to 3 enlarged secondary tubercles adradially. From the sixth plate, secondary tubercles become smaller and fewer in number (1 or 2 tubercles), whereas a secondary tubercle is invariably positioned in the middle of the adradial margin. On the apical surface, each plate carries a secondary tubercle adradially, except for the plate immediately adjacent to the apical opening. In larger specimens (c. 50 mm test diameter), an additional secondary tubercle occurs occasionally ( Fig. 2 A). The size of the primary tubercles decreases rapidly adapically, accompanied by an increase in size of the secondary tubercles after the seventh or eighth plate. On the apical side, there is a distinct, naked V-shaped interradial area, which develops after the seventh to eighth plate. These naked zones are slightly depressed, especially that of the posterior interambulacrum.

Spines. The primary spines are long, slender and smooth, their length equalling about twice the test diameter. The acetabulum is crenulate, the base is only short. A well-developed milled ring is present, which shows a fine striation. The spines are straight to slightly curved ( Fig. 4 B). The collar is short (about 4 mm long at 84 mm length of the spine). The spines are slightly swollen and strongly facetted close to the collar ( Fig. 4 C), this facetted zone being pentagonal or hexagonal in cross-section ( Fig. 4 D). The spines taper to a pointed or faintly spatulated tip.

Discussion. Although usage of the name Phymosoma granulosum (sensu Cotteau, 1864) has been almost consistent and well established in the literature (see Smith & Wright 1996: 286), it in fact refers to a species which differs from the ‘Maastricht’ specimen of Goldfuss (1829), as previously noted by Geys & Jagt (1986) and assumed by Kutscher (1985). Unfortunately, the surviving type specimen is incomplete and some very important test features are missing. However, the unique combination of features displayed by this specimen (i.e., sunken peristome, presence of feeble phyllodes, enlarged secondary adradial interambulacral tubercles on the oral surface) are otherwise observed exclusively in specimens from the lower Maastrichtian white chalk of the Isle of Rügen and a few other areas in northwest Europe. Additionally, the mamelons are of comparable sizes in specimens from Rügen and the ‘Maastricht’specimen of Goldfuss. It is therefore very likely that material from Rügen is conspecific with Cidarites granulosus Goldfuss, 1829 .

Geys & Jagt (1986) were of the opinion of that Cidarites corollaris Leske, 1778 and Cidarites granulosus should be synonymised, as based on the description of Cyphosoma corollare supplied by Cotteau (1864: 669, pl. 1165). Leske’s Cidarites corollaris is based on an internal mould of flint; Smith & Wright (1996: 294) rightly considered this a nomen nudum, a view adopted here. Specimens described as Cyphosoma corollare by Cotteau show certain similarities to P. granulosum , as interpreted here, i.e., the faintly sunken peristome and the feebly developed biserial pore zones adapically. However, Smith & Wright (1996) reassigned those specimens to Gauthieria spatulifera (Forbes in Dixon, 1850); we concur because Cotteau’s drawings of Cyphosoma corollare (1864: pl. 1165) show features (no development of phyllodes, a moderately sunken peristome, an irregular biserial pore zone arrangement adapically, secondary tubercles confined to the oral side, small mamelons and prominent interradial and adradial miliary bands), which are present in the lectotype (subsequent designation of Smith & Wright 1996) and other specimens of G. spatulifera described in Smith & Wright (1996).

Smith & Wright (1996) synonymised Rachiosoma granulosa (Goldfuss, 1829) , as described by Kutscher (1985) from the lower Maastrichtian of Rügen, with Gauthieria middletoni ( Woodward, 1856) . Like C. granulosus , G. m i d d l e t o n i shows a similarly rapid decrease in size of adapical primary tubercles and a similar size of secondary tubercles on the adapical interambulacral plates. However, the latter species differs in having no phyllodes, smaller-sized mamelons and more delicate granules, as well as a smaller and much more sunken peristome. The alternation of simple and polygeminate ambulacral plates next to the peristome, which characterises G. middleton , is not seen in P. granulosum .

Cidarites granulosus has previously been assigned by Kutscher (1985) to the genus Rachiosoma Pomel, 1883 ; this is erroneous. This genus (type species: Cyphosoma delamarrei Deshayes in Agassiz & Desor, 1846) is not a phymosomatid, as it lacks phymosomatoid ambulacral plating. It is better assigned to the family Glyphocyphidae Duncan, 1889 (see Smith, 2011). Correct generic assignment of Cidarites granulosus is fraught with difficulties, because the species shows a mix of features held typical of various phymosomatid genera.

Cidarites granulosus differs from the type species of Phymosoma , Cidaris koenigii Mantell, 1822 , in having a weakly, yet distinctly, sunken peristome; in C. koenigii , the peristome is flush to the test. In contrast, in the type of the genus Gauthieria Lambert, 1888 , Cyphosoma radiatum Sorignet, 1850 , the peristome is deeply invaginated. The pore pairs in C. granulosus are horizontally oriented to slightly oblique and the pore zones expand to form short phyllodes adorally, features in common with P. k o en ig i i. The pore pairs in G. radiata , however, are strongly oblique to near-vertical and phyllodes are never developed. Another difference is found in the structure of the ambulacra: quadrigeminate (or trigeminate) plates alternate with simple elements in G. radiata , whereas in P. koenigii and C. granulosus there are no simple elements intercalated between any polygeminate plates except near the apex. The adapical pore zones are strictly biserial in larger-sized specimens of P. koenigii , whilst in C. granulosus those zones, if present, show only some irregular development of biseriality. In G. radiata , in contrast, these areas are strictly uniserial. Finally, the mamelons of the primary tubercles are intermediate in size in Cidarites granulosus , between the large mamelons of Phymosoma koenigii and the very small ones in G. radiata .

Smith & Wright (1996) noted additional criteria to distinguish species of Phymosoma and Gauthieria , on the basis of British material. Spines in species of Phymosoma (where known) are short and stout, less than the test diameter in length, while those of species of Gauthieria are rather long and slender. While this is true of the type species, spines are known in very few other species, making this distinctive feature less reliable. Another point is that crenulation of the tubercles is often more strongly developed in species of Phymosoma than in species of Gauthieria . In terms of crenulation, Cidarites granulosus is closer to species assigned to Phymosoma .

The type species (by subsequent designation of Smith & Wright, 1996) of Cosmocyphus , Cyphosoma saemanni Coquand, 1860 , is in all respects comparable with Gauthieria radiata , except for the biserial pore zones adapically in the former.

In conclusion, it would appear that there is a larger overlap of characters between Phymosoma koenigii and C. granulosus than between Gauthieria radiata and C. granulosus . Accordingly, C. granulosus is here included in the genus Phymosoma rather than in Gauthieria .

koenigii ) and Gauthieria (i.e., Cyphosoma radiatum ) with Cidarites granulosus , here referred to as Phymosoma granulosum .

Phymosoma granulosum — Desor 1855: 87.—Wright 1870: 129, pl. 23, fig. 2.— Ravn 1928: 55 (pars), pl. 6, fig. 3.— Geys 1980: 204, figs. 1/7, 2/1–4.— Smith & Wright 1996: 286, text-fig. 100; pl. 99, figs. 1–6.— Jeffery 1997: 672, text-figs. 5a, b; pl. 3, figs. 1–5.— Smith et al. 1999: 96, pl. 2, figs. 3, 4.— Smith & Jeffery 2000: 86, text-fig. 32 D, E.

Phymosoma gr. granulosum — Jagt 2000: 229, pl. 9, figs. 1–6; pl. 10, figs. 13, 14. Cyphosoma granulosum — Woodward 1856: 1.—Cotteau 1864: 685, pl. 1169, figs. 1–8. Phymosoma tiara — Lambert 1911: 67 (pars).

Phymosoma corneti — Geys 1980: 200 (pars), figs. 1/1–5 only.

Phymosoma koenigi — Kutscher 1985: 521 (pars), pl. 2, figs. 1–3.

Rachiosoma corollare sensu Cotteau, 1865 — Geys & Jagt 1986: 99, figs. 3a–i, 4a, b. Phymosoma sp. van der Ham 1988: 140, pl. 4, figs. 1–4.

Types. Holotype is NHM E1934, from the upper Campanian of Norwich, Norfolk; paratypes are GZG.STR.018614 and 0 18816 (leg. M. Kutscher) from the lower Maastrichtian, Promoisel quarry, Rügen.

Stratum typicum. Upper Chalk, Campanian.

Locus typicus. Norwich, Norfolk, southeast England.

Derivatio nominis. In honour of the Danish palaeontologist, J. P. J. Ravn.

Material. In addition to the type material: MB.E 6566 (locality unknown, probably Rügen), GZG.STR.018629, lower Maastrichtian, Promoisel Quarry, Rügen, Germany.

Occurrence. Lower Campanian of the Hannover area, northern Germany (U. Frerichs Collection, Langenhagen, Germany); upper Campanian of Norfolk, southeast England; lower Maastrichtian of Rügen; lower to upper Maastrichtian of Møns Klint, Stevns Klint and Aalborg, Denmark; lower to middle Danian Bryozoan Chalk, Stevns Klint and Fakse, Denmark; upper Danian Crania Chalk, Herfølge , Denmark ( Ravn 1928); upper Maastrichtian and Danian of Mangyshlak, Kazakhstan ( Jeffery 1997); lower to middle Danian, Maastricht area, the Netherlands and Belgium; upper Campanian and upper Maastrichtian of Mons Basin (southern Belgium) ( Geys 1980; Geys & Jagt 1986; van der Ham 1988; Jagt 2000), Maastrichtian of Orglandes (Manche, northwest France; see Lambert 1909); lower and upper Campanian of the Miechów area, southern Poland ( JWMJ, pers. obs.); Maastrichtian of Santander, Cantabria, northern Spain (NHM). See also the list of occurrences in Smith & Jeffery (2000: 86).

Diagnosis. Large species of Phymosoma with a flush peristome and a very large pentagonal apical opening, which projects markedly into the posterior interambulacrum; very well-developed biserial pore zones adapically and well-developed phyllodes present adorally; tuberculation and granulation coarse; only very small secondary tubercles present adradially on interambulacral plates above ambitus.

Description. For a very detailed and comprehensive description of this species, reference is made to Smith & Wright (1996: 286). Here we present a synopsis and add some data on spine morphology. Large tests (45 mm), flat in profile (height equalling 36–45 per cent of test diameter), circular to subpentagonal in outline. The tubercles are crenulate and imperforate. The pentagonal apical opening is very large (length equalling 46 per cent of test diameter).

Peristome. Flush, occupying 36–43 per cent of test diameter. Buccal slits are faintly developed. Lantern elements (hemipyramids and keeled teeth) are preserved in GZG.STR.018616, yet unfortunately mostly covered by the test ( Fig. 5 E). Other records are so far unknown.

Ambulacra. The ambulacral zones are 23–25 per cent of test diameter at ambitus, compounded in phymosomatoid style. At a diameter of 41.8 mm, there are 11 tubercles in a row. The ambulacra taper adapically and, to a lesser extent, adorally. Phyllodes are very well developed adorally. Adapically, pore zones are arranged biserially up to the apical opening, (formed after 33–37 pore pairs). Pore pairs are somewhat oblique in position at and below the ambitus, above they are more or less horizontal. At the ambitus, there are six pore pairs per plate. The areoles in a column are confluent subambitally, there is some variation due to size; in smaller-sized individuals they may be separated by small rows of granules (see Smith & Wright 1996). Adorally, tubercles decrease gradually in size, while adapically there is an abrupt decrease in size of the tubercles above the eighth to ninth plate. The perradius is characterised by a small double band of granules and small tubercles.

Interambulacra. The interambulacra are about 34 per cent of test diameter at ambitus. There are 10 or 11 tubercles in a column at a test diameter of 45 mm. The mamelons are 37–40 per cent of the areole diameter. There is some variation in the degree of confluence of the areoles: either becoming confluent at the ambitus or only below it. A broad, somewhat depressed, V-shaped naked zone is present at the interradius on the apical surface. Below, the interradius has a well-developed band of granules and small tubercles, as well at the adradius. At the oral surface, the first 4 to 5 plates have 2 to 4 secondary tubercles adradially. Above the ambitus, very small secondary tubercles occur only sporadically; these are seen first at test sizes of about 45 mm (see Smith & Wright 1996).

Spines. The primary spines are rather stout, cylindrical and circular in cross-section. The acetabulum is crenulate. The finely striate milled ring, above the short button, is well developed. The short collar has a fine striation. The tips of the spines are gently curved and flattened. However, a specimen from the upper Maastrichtian of Stevns Klint, Denmark (Danekrae no. 154, housed at the Geologisk Museum, København, Denmark), with a large number of spines closely associated with the test, has recently been examined by one of us (MK); this is potentially conspecific with P. r a v n i sp. nov. The spines are slightly curved and about twice as long as the test diameter; however, because only the oral side of the test can be seen, definitive placement of this specimen is currently impossible.

Discussion. Phymosoma ravni sp. nov. is very similar to P. koenigii in the size of the mamelons, the timing of the development of biserial pore zones and the number of pore pairs in ambital ambulacral plates. It differs from the latter only by its larger apical opening and the paucity and feeble development of secondary interambulacral tubercles adapically; in P. koenigii these tubercles are very much larger (for a more detailed comparison see Smith & Wright 1996).

Lambert (1911: 67) grouped several species under the name Phymosoma tiara (Agassiz in Agassiz & Desor, 1846), regarding them as forms of P. tiara , including Cidarites granulosus , Cidaris koenigii and Cyphosoma granulosum (sensu Cotteau 1864) [= P. r a v n i sp. nov.]. The latter (form C) Lambert stated (p. 70) as corresponding to the type of P. tiara (M6, Agassiz Collection, Neuchâtel, Switzerland). However, there are clear differences between the type of P. t ia r a and C. granulosum (sensu Cotteau) . Phymosoma tiara is distinguished from the latter in having a faintly sunken peristome, presumably uniserial pore zones adorally and a widely spaced granular area between the ambulacral tubercles adorally. From this it would appear that P. t i a r a should more correctly be assigned either to Gauthieria or Cosmocyphus .

Lambert (1911) delineated form D (p. 70) from form C by its more depressed test, more closely spaced ambulacral tubercles next to the apex and comparatively small and scattered secondary tubercles, features also shown by Phymosoma ravni sp. nov. Lambert (1911: 70) named this form (form D) Phymosoma normannum , on the basis of specimen M56 (a mould, Agassiz Collection, Neuchâtel). This mould is insufficiently well preserved to show any details of pore arrangement and the whereabouts of the original specimen are unknown. Accordingly, the name Phymosoma normannum is here treated as a nomen nudum.

On the basis of a test fragment from Danian (early Paleocene) strata, Ravn (1928: 57, pl. 6, fig. 5) erected Phymosoma subconicum ; this shows neither the apical opening nor the peristome. Ravn characterised it as very inflated (conical) and with biserial pore zones adapically. In Ravn’s ‘species’ the tubercles seem to be smaller and less prominent than in our new form. In view of the very poor state of preservation of the type specimen, we here treat Phymosoma subconicum as a nomen dubium until further material from the type locality, Herfølge, Denmark, becomes available.

Figures

FIGURE 1. Camera-lucida drawings of plating style in Phymosoma granulosum (Goldfuss, 1829), based on GZG. STR. 018536 from the lower Maastrichtian of Wittenfelde Quarry, Rügen, northeast Germany (see also Fig. 3 A – C); A, interambulacral zone; B, ambulacral zone. Scale bar equals 2 mm.

FIGURE 2. Phymosoma granulosum (Goldfuss, 1829), lectotype (RFWUIP 1335), of early to early late Maastrichtian age, locality unknown; A, oral view; B, ambital view; C, oral detail; D, detail of ambitus. A, B, natural size x 1,6; scale bar in C, D equals 2 mm.

FIGURE 4. Phymosoma ravni sp. nov.; A, GZG. STR. 018629, lower Maastrichtian, Promoisel Quarry, Rügen, Germany, test with spines. Phymosoma granulosum (Goldfuss, 1829); B, GZG. STR. 018585, lower Maastrichtian, Wittenfelde Quarry, Rügen, Germany, accumulation of spines; C, D, MB. E 6867, lower Maastrichtian (Trigonosemus pulchellus Zone), Promoisel Quarry, Rügen, Germany, lower third of spine and spine in cross-section, respectively. All specimens coated with ammonium chloride prior to photography; natural size (B), x 1.2 (A), x 2 (C) and x 3 (D).