? Haplacanthus sp. – Lelièvre 1988: 300, fig. 5.
? Phoebodus sp. – Forey et al. 1992: 31, fig. 7.
Nogueralepis sp. – Burrow et al. 2006: 554, fig. 11: 12- 17, 23, 24.
“Chondrichthyan teeth, scales” – Derycke & Goujet 2009: 87.
Tassiliodus lessardi “nomen nudum”– Derycke & Goujet 2010: 145.
TYPE LOCALITY. — Near In Guezzam, Oued Felaou or Anou Izileg, southern Ahaggar (Algeria).
ETYMOLOGY. — After L. Lessard, a hydrogeologist who discovered the material in the 1960’s.
DIAGNOSIS. — As for genus.
SYNTYPES. — Teeth, MNHN.F.ALD - 15 (Fig. 4A) and MNHN.F.ALD-16 (Fig. 4B); scales MNHN.F.ALD-17 to 35; thin sections MNHN.F.ALD-30-32, 34, 35.
OTHER MATERIAL EXAMINED. — 107 scales, same origin as the syntypes (see Material and methods).
DESCRIPTION OF THE TEETH
The tooth crown of MNHN.F.ALD-15 (Fig. 4A) bears 8 cusps and cusplets. Among the three main cusps, the middle one is half the height of the lateral cusps, that diverge at an angle of about 50°. The main cusps have a rounded tip. Five cusplets are visible, three on one side of the central cusp and two on the other side. Lateral cusplets are not in the same plane, and the smallest ones are in a more labial position. The smooth crown is probably covered with enameloid, and only one ridge is visible in the middle of a main cusp (right side in Fig. 4A 1).
The base is perpendicular to the crown and expanded lingually. Its internal structure is revealed by a fracture surface showing bone or osteodentine (Fig. 4A 3) with a rugose external aspect (Fig. 4A 4). There is apparently no regular organization of the foramina in the base (Fig. 4A 4, A 5).
The crown of the second, smaller tooth MNHN.F.ALD-16 (Fig. 4B) is only visible in lingual aspect and made up by four cusps with 3 diverging main cusps (angle around 50°). The tip of the cusps are more pointed and slender than in MNHN.F.ALD-15. The rugose base shows a kind of lingual bulge behind the central cusp (Fig. 4B 4). One cusplet is visible but is not in the same plane, it is situated more labially, like in MNHN.F.ALD-15. The base is poorly developed and the cusps are oriented or curved lingually.Tiny foramina are visible in the lingual part of the base.
Remarks
Our specimens have been compared with other Emsian chondrichthyans: Doliodus Traquair, 1893, known from an articulated specimen (Turner & Miller 2005: fig. 6) as well as isolated teeth, Protodus Woodward, 1892 (Turner & Miller 2008), Pucapampella Janvier & Suarez-Riglos, 1986 (Maisey & Anderson 2001), and possible Emsian Antarctilamna Young, 1982 (e.g., Forey et al. 1992).
Devonian levels of Saudi Arabia have yielded early vertebrate material. A tooth referred to Phoebodus sp. (Forey et al. 1992: fig. 7) has been described from the Jauf Formation (Saudi Arabia) and dated as Pragian to Early Emsian. Although it is only illustrated by a drawing and in lingual aspect, it is possibly an antarctilamnid, a family mainly known from the Givetian-Frasnian (type genus Antarctilamna). However this tooth appears smooth and shows fewer vascular foramina than Antarctilamna and is not comparable to Tassiliodus n. gen. which shows no big foramina. Antarctilamna is also represented by spines and scales in the Wajid Formation (Saudi Arabia, Pragian to Early Emsian) (Forey et al. 1992). Other “shark” material corresponds to “ Cladodus ” teeth, that are not identifiable according to Ginter (2004). One species, A. seriponensis, was also described from the Emsian of Bolivia (Gagnier et al. 1988) based on fin spines. Antarctilamna teeth are characterized by many ridges on the cusps (Young 1982; Long & Young 1995), a feature that is absent in our specimens.
Pucapampella first found in Bolivia (Janvier & Suarez-Riglos 1986) is also known from the upper Emsian in South Africa (Anderson et al. 1999; Maisey & Anderson 2001 in Maisey 2004). It is only known by its braincase and no teeth have been yet described (Maisey 2001).
Doliodus problematicus Woodward, 1892, is early Emsian in age (Canada) and shows trunk scales (Miller et al. 2003: 503) that resemble those of Antarctilamna from the Jauf Formation. Doliodus teeth have cusps similar to some xenacanthiform ones and have been included in the Omalodontida (Miller et al. 2003). Known as one of the oldest articulated chondrichthyans, Doliodus possesses paired fin-spines, a character long considered as a synapomorphy of acanthodians, but now as a gnathostome symplesiomorphy (Miller et al. 2003). Turner (2004) described the teeth of Doliodus in detail. The main difference with Tassiliodus n. gen. concerns their base, which is oriented labially in Doliodus, hence its inclusion in the Omalodontida. The crowns are also different: in Doliodus one of the main lateral cusps is larger than the other, the proportions in length of cusps are different, and intermediate cusplets are in the same plane as the main cusps.
Protodus jexi Woodward, 1892 from New Brunswick (Canada) (late Pragian/early Emsian) is quite different, as it has monocuspid teeth (Turner & Miller 2008).
Finally among Emsian chondrichthyans, the Tassiliodus n. gen. teeth show a unique morphology.
Leonodus Mader, 1986, first described by Mader (1986) from Lochkovian-Pragian levels of Spain, is considered as the oldest unequivocal chondrichthyan teeth (Turner 2004; Turner & Miller 2005). It is bicuspid, more acute when sharpened by wear (Botella et al. 2009), and is the first indication of a dental lamina (Botella 2006, except for placoderms according to Smith & Johanson 2003, even there is a regular tooth addition without a true dental lamina, M. M. Smith pers. comm.).
DESCRIPTION OF THE SCALES
From the c. 126 scales, six morphotypes have been identified as belonging to the squamation of Tassiliodus lessardi n. gen., n. sp.
High-neck scales, almost 58 scales (Fig. 5 A-D) These four scales share a crown that shows wide grooves intercalated with converging thin, sometimes nodose, ridges in the middle, an elongate neck and a flat base, that is narrower than the crown. These scales are generally higher than wide.
The crown of MNHN.F.ALD-17 (Fig. 5A) bears six main tuberculated ridges converging at the top. One ridge is divided near the crown margin (right on Fig. 5A 1). One intermediate ridge seems to merge on the opposite side. This round scale has a high neck and a flat base with one basal pore and a notch near its margin (Fig. 5A 2). In the scale MNHN.F.ALD-18 (Fig.5B), more numerous ridges begin in the vertical part of the crown, converging at the top, nipped-in waist and high neck with one foramen at the extreme basal part of the neck.
In the scale MNHN.F.ALD-19 (Fig. 5C), one foramen in the middle of the base and one tiny lateral foramen, could correspond in position to the “pulp” cavity “openings” described in Elegestolepis grossi Karatajute-Talimaa, 1973 (Karatajute-Talimaa 1973: taf. 3), but in our scale there is no pulp cavity, so they are strictly vascular foramina.
A bigger scale, MNHN.F.ALD-20 (Fig. 5D), shows about nine ridges converging and reaching the top and slightly displaced. Smaller ridges starting at the margin of the crown are intercalated between them. The crown outline is oval and the margin crenulated. The elliptical base has the same orientation as the crown. In basal view (Fig. 5D 2), a tenuous limit underlines the boundary between the crown and the neck.
Flat-crowned scales (Fig. 5 E-G)
The scales MNHN.F.ALD-21 to 23 are flatter. The upper crown is ornamented with 4-7 ridges converging backward and separated by wide grooves. The lower lateral crown is either smooth (Fig. 5E 1), or with ridges parallel or perpendicular (Fig. 5F 1) to the central crown ridges. The flat base, narrower than the crown, may show two foramina in MNHN.F.ALD-22 (Fig. 5F 2): one near the centre and one near the margin.
The morphology of the scales is similar to that of Wetteldorfia triangula Vieth-Schreiner, 1983 (Vieth-Schreiner 1983: taf. 1. 1-9; Emsian-lower Eifelian), which is considered as an acanthodian. Here the circular base is smaller than the crown whereas Wetteldorfia Vieth-Schreiner, 1983 shows a forwardly displaced, diamond-shaped base. The scales are smaller (1-1.5 mm) than those of Wetteldorfia (3 mm). The histological structure of some scales of Wetteldorfia from the Eifel (ViethSchreiner 1983: abb. 7, 8) recalls that of the highneck scales (Figs 5 A-D; 7E 2), with a box-in-box structure with only two or three growth lines in the crown, which do not continue in the base. In Vieth-Schreiner’s drawings, internal growth zones seem to terminate before the surface. Furthermore, the same tissues are visible in the same position. Usually, in the center of the crown of acanthodian scales, a primordium is visible. It is present neither in the thin sections of Wetteldorfia from the Eifel, nor in our material (morphologically more closely corresponding to Figure 5 A-D; the material that shows the external morphology of Wetteldorfia has not been sectioned). Wetteldorfia is also described from the Emsian of Saudi Arabia (Burrow et al. 2006: figs 4.12, 13, 15, 16) but is different from the type species, and Burrow et al. (2006: 544) proposed a revised diagnosis of the genus.
Leaf-shaped scales (Fig. 6A, B)
These flat diamond-shaped scales show ridges converging in the centre of the convex crown, some very worn (ALD-25; Figs 6B 1; 7B 2) and others less so (ALD-24; Fig. 6A 1). Shorter ridges appear on the lateral sides of the crown (Fig. 6A 2, B 2). The flame outline corresponds to two concave opposite sides and two other more convex ones (Fig. 6A 1). The base is flat (Fig. 6A 2) or slightly convex (Fig. 6B 2). The neck is short and nipped-in waist (Fig. 6A 2). These scales are wider than high.
Spiny scale (Fig. 6F)
A spiny crowned scale, MNHN.F.ALD-29, has a nipped-in waist, high neck and a flat base (Fig. 6F 2). Three broad spines are distributed with one at the top of the crown and two on one side (Fig. 6F 1). Smaller, more irregularly distributed, spines may be joined by a crest to the highest spine in two files (Fig. 6F 1). It resembles isolated tubercles of the scale MNHN.F.ALD-17 (Fig. 5A 1).
Scales with parallel ridges on the crown (Fig. 6C, D) The two scales MNHN.F.ALD-26, 27 have a flat, posteriorly elongated crown. The crown is ornamented with four, strong, subparallel ridges – the biggest one in the middle – separated by wide grooves (Fig. 6C 1). The second scale MNHN.F.ALD-27 shows seven ridges beginning in the anterior part, and two supplementary ones intercalated in the posterior part (Fig. 6D 2). The bases are different in these two scales: more flat for MNHN.F.ALD-26 (Fig. 6C 2) and more convex and diamond-shaped for MNHN.F.ALD-27 (Fig. 6D 1) with a foramen in a lateral position.
The scale MNHN.F.ALD-26 (Fig. 6C 1) evokes a scale of Elegestolepis grossi Karatajute-Talimaa, 1973, from the Silurian of Tuva (Karatajute-Talimaa 1973: taf. 3, figs 5c, 6a), with a high neck and a flat base, and is associated with scales showing a more diamond-shaped base,like that in MNHN.F.ALD-27 (Fig. 6D 1).
Broken scale (Fig. 6E)
The broken scale MNHN.F.ALD-28 shows a rounded convex base protruding forward of the crown (Fig. 6E 1). One foramen is visible on the high neck. The fore-crown displays ridges that are inserted low on the base.
SCALES HISTOLOGY
Thin section 1 MNHN.F.ALD-30 (Fig. 7A 2), thin section 3 MNHN.F.ALD-32 (Fig. 7C 2) and thin section 5 MNHN.F.ALD-35 (Fig. 7H 2) are vertical sections of scales with a high neck, converging crown ridges and a flat base (Fig. 5 A-D). Thin sections 2 MNHN.F.ALD-31 (Fig. 7B 1) and 4 MNHN.F.ALD-34 (Fig. 7E 2) are of flatter scales like MNHN.F.ALD-24 and 25 (Fig. 6A, B). Thin section 4 MNHN.F.ALD-34 (Fig. 7E 2) is of a scale that is wider than high, but also with wide grooves and thin ridges converging at an eccentric apex. In the Figure 7, each thin section corresponds to the nearest scale morphology and bears the same letter.
Almost the entire scale internal structure comprises a network of meandering canals emerging in any part of the scale; canals can open out on the base (Figs 7A 2, C 2; 5A 2, C 1), the neck (Figs 7C 2; 5B 1) or on the crown (Fig. 7C 2, H 1, H 2). It consists of a distinctive type of mesodentine (Ørvig 1967).
A kind of box-in-box pattern is visible in the crown, but some elements of the previous generations of the scale may appear at the surface. Whereas thin section 3 MNHN.F.ALD-32 (Fig. 7C 2) only shows one growth zone, thin sections 1, 2 and 5 show two growth stages (Fig. 7A 2, B 1, H 2), with possibly more for thin section 4 MNHN.F.ALD-34 (Fig. 7E 2), based on the undulating ridges. The first stage, mainly invaded by the vascular canal network, emerges on the right side on MNHN.F.ALD-35 (Fig. 7H 2) and on the left one in MNHN.F.ALD-34 (Fig. 7E 2). It may correspond to the intermediate external shorter ridges at the crown base (Fig. 7E 1). It means that the overlapping by the second growth stage is not complete and that this histology does not exactly correspond to a real box-in-box pattern.
Resorption-remobilization phenomena are observed:on the left side of the tip of MNHN.F.ALD-30 (Fig. 7A 2), in the middle part and right side of MNHN.F.ALD-34 (Fig. 7E 2) and in the middle bulge of the first stage MNHN.F.ALD-35 (Fig. 7H 2). Resorption was never found before in chondrichthyans, except in Ctenacanthus sp. from the Namurian (Derycke-Khatir 2005).
More classically, bushes of pallial dentine (Radinsky 1961) radiate from vascular canals under the crown surface, sometimes lined by a thin refringent layer. Two levels of dentine may appear inside the crown, for example in thin sections 1 MNHN.F.ALD-30, 4 MNHN.F.ALD-34 and 5 MNHN.F.ALD-35 (Fig. 7A 2, E 2, H 2). Sharpey’s fibers and osteocytes are localized in the extreme basal part of the scales, and may be particularly dense in thin sections 1 MNHN.F.ALD-30 and 5 MNHN.F.ALD-35 (Fig. 7A 2, H2). In thin section 4 MNHN.F.ALD-34 (Fig. 7E 2) the osteocytes are not randomly organized, surrounding vascular canals. The density and the organization of Sharpey’s fibers and osteocytes are variable, possibly reflecting the maturity of the scale.
In two thin sections (1 MNHN.F.ALD-30 and 2 MNHN.F.ALD-31, Fig. 7A 2, B 1), a short canal joining an aperture in the posterior part and in the basal part of the base corresponds to the openings and the notch visible on MNHN.F.ALD-17 (Fig. 5A 2).
Scale histology appears almost identical in the material of Algeria and in the “ Nogueralepis sp.” scales from the Jauf Formation of Saudi Arabia (Burrow et al. 2006: fig. 11: 12-17, 23, 24). Crown ridges converge backwards or are subparallel in “ Nogueralepis sp.”, and the scales have the same morphology as those of Tassiliodus lessardi n. gen., n. sp., with a very high neck and a flat base. They strickingly differ from those of the type-species of Nogueralepis Wang, 1993 from the Lochkovian of Spain (Wang 1993), particularly in their histology. The scales from Saudi Arabia were probably misidentified and may in fact belong to our new taxon Tassiliodus n. gen., reinforcing the close biogeographic affinity between these two areas.
Thin section 4 MNHN.F.ALD-34 (Fig. 7E 2) evokes a thin section of a tiny adult scale of Elegestolepis grossi from the Silurian of Tuva (KaratajuteTalimaa 1973: 42, abb. 3, E), but also the thin section of Wetteldorfia from Saudi Arabia, except for the growth zones of the base (Burrow et al. 2006: fig. 4: 16).
CONCLUSION FOR CHONDRICHTHYAN SCALES Scales with the morphology shown by MNHN.F. ALD-17, 19 (Fig. 5 A-D) and histology with the network of meandering canals as in MNHN.F.ALD-30, 32 (Fig. 7A 2, C 2) resemble those of “ Nogueralepis sp.” from Saudi Arabia (Burrow et al. 2006: fig. 11: 12-17, 23, 24). The scales MNHN.F.ALD-21-23 (Fig. 5 E-G) morphologically resemble Wetteldorfia scales (Vieth-Schreiner 1983: taf. 1.1-9), but the different morphotype of scales MNHN.F.ALD-17-19 (Fig. 5 A-D) corresponds to the histology of Wetteldorfia (Fig. 7E 1, E 2). Thin section MNHN.F.ALD-34 (Fig. 7E 2) also evokes the histology of Elegestolepis, except for the lack of a pulp cavity and its finer calibre canals (Karatajute-Talimaa 1973: 42, abb. 3, E). Two other scales, MNHN.F.ALD-26, 27 (Fig. 6C, D), resemble the morphology of Elegestolepis (Karatajute-Talimaa 1973: taf. 3, figs 5c, 6a).
The distinctive histological structure indicates that all these scales may belong to the squamation of the same chondrichthyan.