Introduction

Palaeozoic echinoderms produced a wealth of unusual thecal morphologies, which are hard to interpret in terms of taxonomic assignment and/ or functional morphology. Although this is particularly true for Early Palaeozoic clades (e.g., Sumrall and Wray 2007), it still holds true for some Devonian forms, for example pyrgocystid edrioasteroids (e.g., Bather 1915; Dehm 1961; Holloway and Jell 1983; Klug et al. 2008; De Baets et al. 2010), some crinoids (e.g., Schmidt 1934; Bohatý 2011) or late blastozoans (e.g., Dehm 1934; Rahman and Lintz 2012). Crinoids show their highest disparity in the Ordovician, but disparity remains high until the end of the Devonian (Foote 1995). Several areas of the Palaeozoic crinoid morphospace have not been reoccupied in the Mesozoic (Foote 1999).

In recent years, we have discovered the first two specimens of the genus Tiaracrinus in Morocco. This genus is usually thought to be restricted to Early (Lochkovian) to Middle Devonian (Eifelian; see Le Menn 1987; Prokop 1987; Haude 2008), but Kříž (1992) reported some in open nomenclature from the Late Silurian (Ludfordian,). Tiaracrinus is peculiar for the following reasons: (i) the calyx shows a tetramerous symmetry; (ii) so far, it has not been found with arms, although some unusual arm structures have been assigned to the genus (see further); (iii) the theca has four fields with epispires, giving it a blastozoan appearance.

Species of Tiaracrinus based on calices were first described in the 19^th century. Schultze (1866) introduced the genus with the type species T. quadrifrons in his monograph on the Middle Devonian echinoderms from the Eifel Mountains. Soon thereafter, two more species of Tiaracrinus were described, T. oehlerti Schlüter, 1881 and T. rarus (Barrande, 1887). In the 20^th century, two additional species were published based on calices, T. tedraedra Jaekel, 1901 and T. moravicus Ubaghs and Bouček, 1962. Two species based on limbrachoids, unusual arm structures, have been recently assigned to Tiaracrinus with some reservation (T.? obtusibrachialis Haude, 1993 and T.? aceribrachialis Haude, 1993). Hauser (2008: 1) listed the latter two species as nomina dubia. Additional specimens in open nomenclature were reported from the Ludfordian, Lockovian, Pragian and Emsian of the Czech Republic (Prokop 1987; Kříž 1992; Prokop and Petr 2002).

Fig. 1.

Geological map (modified from Klug 2002) of the eastern Anti-Atlas showing the two localities that yielded Tiaracrinus.

In this study, we describe and illustrate the first two specimens of Tiaracrinus from the Early Devonian of Morocco. We discuss ontogenetic changes within this group as well as its phylogeny.

Institutional abbreviations.—GFCL, collection of D. Le Maître, Laboratoire de Géologie de la Faculté Catholique de Lille, France; LPB, Laboratoire de Paléontologie de Brest, France; PIMUZ, Paläontologisches Institut und Museum, Universität Zürich, Switzerland.

Other abbreviations.—PCA, Principal Component Analysis.

Material

So far, only two specimens of Tiaracrinus have been found in the Moroccan eastern Anti-Atlas by us. Both specimens are of latest Emsian age and were found at the southern edge of Jebel Oufatene (PIMUZ 29741) and at the “red cliff” in Hamar Laghdad (PIMUZ 29739; Fig. 1). Both specimens are stored at the Paläontologisches Institut und Museum at the University of Zurich. Both specimens (Fig. 2) are slightly corroded but well preserved otherwise. The specimen from Hamar Laghdad is slightly better preserved, probably due to a partial silicification. Holdfast, stem, brachials, arms and oral plates are not preserved in both specimens.

Fig. 2.

Zophocrinid crinoid Tiaracrinus jeanlemenni sp. nov., probably late Polygnathus patulus Conodont Biozone, late Anarcestes lateseptatus Ammonoid Biozone, latest Emsian, Early Devonian, eastern Anti-Atlas Morocco. A. PIMUZ 29739, holotype, “Red cliff” at Hamar Laghdad, Tafilalt. Lateral views (A₁, A₂), showing the radial channel and the rib-fields; note the ornamentation in the channel near the oral surface in A₂,. Aboral view (A₃), note the small cross section of the trimeral basals and the low rim around it. Oral view (A₄), note the ornamentation in the channel near the oral surface. Detail of A₃ (A₅), note the epispires and the uniform thickness of the ribs. On the top right, the subtle striation is faintly visible between the ribs. Oblique view of the aboral side (A₆), showing the epispires. Due to weathering, the ribs obtained a tuberculate ornament (A₇). B. PIMUZ 29741, paratype, Jebel Oufatene, Maïder. Aboral view (B₁), the rib-fields are less vaulted and the cross section of the calyx less quadrate than in the holotype; the fossil on the top left is a spiriferid. Lateral views (B₂, B₃), showing the narrow ribs and the smooth surface of the radial channels; in the lateral aspect it looks like the radial channel is tapering towards the oral side of the cup.

PIMUZ 29739, the larger specimen, was collected from scree atthe “red cliff” (Klug2002;N 30.82453°, W4.90210°) at Hamar Laghdad in the Tafilalt region. Hamar Laghdad is situated about 18 km ESE of the town Erfoud (Fig. 1). Hamar Laghdad has become world-renowned for its mud-mounds, some of which being completely exhumed, some others are still more or less covered by the overlying sediments (e.g., Roch 1934; Massa et al. 1965; Hollard 1974; Alberti 1982; Brachert et al. 1992; Wendt 1993; Belka 1994,1998;Bultynck and Walliser 2000; Aitken et al. 2002; Berkowski 2006; Cavalazzi et al. 2007). As far as palaeontology is concerned, this locality yielded highly diverse trilobite associations (e.g., Alberti 1969, 1982; Klug et al. 2009), abundant cephalopods (Töneböhn 1991; Klug 2001, 2002; Klug et al. 2009) including ammonoids infested by parasites (De Baets et al. 2011) as well as many blastozoans, crinoids (sometimes encysting corals: Berkowski and Klug 2012), brachiopods, rugose corals specialised on vent habitats (Berkowski 2004, 2006, 2008,2012, Belka and Berkowski 2005), locally favositids in rock-forming numbers and large bivalves (Panenka, Patrocardia and similar forms). The specimen PIMUZ 29739 was collected from marl containing limestone nodules and layers of latest Emsian age. These layers contain ammonoids of the genera Amoenophyllites, Anarcestes, and Chlupacites (Klug et al. 2009). Immediately below these marl beds, a claystone succession of late Emsian age is present, which overly the massive mudmound limestone, also of Emsian age. The latter limestone contains Mimagoniatites fecundus, thus confirming their basal Dalejan age (e.g., Klug et al. 2009). Above the marl, the layers become increasingly limy. At the base of the limestone, a dark nodule layer contains abundant Foordites veniens (Klug 2002; Klug et al. 2009). Diagenesis and taphonomy of the red cliff are discussed in Klug et al. (2009).

Fig. 3.

Bivariate plots with various measurements and ratios of some published (Le Menn 1190; Hauser 2008) and the two newly described specimens of Tiaracrinus. See the methods chapter for definition of the parameters. A. Calyx width vs. calyx height; note the clear sepration of taxa. B. Calyx shape vs. calyx height. C. Number of pore rows vs. calyx height. As in A, the new taxon is also well separated from the existing taxa. The linear trendline probably reflects an ontogenetic trend. Open squares mark the values of the new species T. jeanlemenni sp. nov., solid squares mark all other species.

PIMUZ 29741 is somewhat smaller and was found at the southern edge of Jebel Oufatene (Fig. 1; compare Hollard 1974; Massa et al. 1965). This specimen is slightly deformed, probably due to compaction of the sediment. However, this deformation is minimal. We tested this by taking high and low estimates of the measurements. Therefore, we think that the overall proportions are reasonably well preserved. This is additionally corroborated by comparing the superficial visual similarities and differences of the measured specimens with PIMUZ 29741 with the clustering of forms in the PCA and in the cluster analysis (see the methods chapter). It was collected from a trench where locals had dug for late Emsian trilobites (N 30.824532°, W 4.90210°). Jebel Oufatene is about 14 km S of Fezzou in the Maïder region (eastern Anti-Atlas, Morocco). According to Kaufmann (1998), the entire Emsian is about 100 m thick in this area. The Dalejan (late Emsian) is represented by a thick claystone and clayey marl sequence, which displays increasing carbonate content near its top. For a sedimentological basin model of the Maïder Basin: see Döring (2002).

Methods

In order to distinguish between the species of this genus based on morphological differences, we produced bivariate plots (Fig. 3), a Principal Component Analysis (PCA; Fig. 4) and a Cluster Analysis (Fig. 5) using PAST (Hammer at al. 2001) based on the following six characters (Table 1):

Table 1.

Values used in the principal component and cluster analyses. Most of the values were taken from Hauser (1997,2008). Values with some uncertainty are given in bold.

Calyx width size class (based on the largest width): The largest specimen was set as 1.

Calyx height size class: The largest specimen was set as 1.

Curvature of the calyx walls, calculated by dividing the largest width of the oral surface (assumed to be more or less identical with the diameter of the oral edge of the radial circlet) by the maximum width of the calyx.

Stem to calyx-ratio: The ratio between the diameter of the facet of the most proximal nodal columnal and the diameter the maximum width of the calyx.

Row ratio: The number of epispire rows divided by the maximal number of epispiral rows of all known specimens.

Shape: The values of character 1 (width size class) divided by those of value 2 (height size class).

The standardisation was made for several purposes: (i) standardisation reduces the bias caused by differences in dimensions of the respective parameter; (ii) it eliminates the bias in scaling errors in those cases, where morphometric values were taken from the literature. We thus performed a PCA on the correlation matrix. The intraspecific variability is mainly caused by differences in ontogenetic trajectories (see Fig. 3) and, to a minor degree, by preservational biases such as the differences between internal moulds versus specimens that preserve the skeleton.

In the bivariate plots (Fig. 3), which were made using Excel, trendlines are given, which are supposed to show the predominant ontogenetic trend (correlation with calyx height, i.e., size). These were also produced with Excel.

The choice of specimens is based on their availability and preservation mainly. Not many well-preserved specimens of Tiaracrinus are sufficiently well documented and show all characters needed for our analyses, hence the low number of data.

Fig. 4.

PCA of some published (Le Menn 1190; Hauser 2008) and the two newly described specimens of Tiaracrinus (see Table 1). See Methods for definition of the parameters. A. Plot of principal components 1 and 2; note, how the new species separates well from the previously known ones. B. Plot of principal components 1 and 3. C. Plot of principal components 2 and 3; again, the new species is morphologically separated from the others. Open circles mark the values of the new species T. jeanlemenni sp. nov., solid squares mark all other species.

Systematic palaeontology

Fig. 5.

Cluster analysis (Euclidean, paired group) of some published specimens and the two newly described specimens (see Table 1). The cluster on the right including Tiaracrinus quadrifrons and Tiaracrinus tetraedra is supported by the bootstrap (at 500 replicates; 59%with past); the Tiaracrinus oehlerti and Tiaracrinus moravicus-group (59 and 64%) and the Tiaracrinus jeanlemenni and Tiaracrinus rarus-group (62 and 72%) are also reasonably well supported. See also the methods chapter for definition of the parameters.

Growth

As already pointed out by Le Menn (1990: 165), some of the characters of Tiaracrinus, such as the number of ribs, vary through ontogeny. The correlation of calyx width and height in the bivariate graph of Fig. 3 A is not significant (r = 0.358, R² = 0.128, p = 0.1897), which is probably due to the fact that the graph depicts specimens from several species. Evaluation of a possible correlation with single species is, however, not possible at this point because of the lack of material and thus data. In spite of the lack of data, it is still clear that with increasing height, the width is also increasing, apparently to differing degrees in different species, but this hypothesis requires much more material to be tested. The number of ribs and epispires increases more or less constantly through ontogeny (Fig. 3B; r = 0.0695, R² = 0.2213, p = 0.4809), namely from 7 to 10 in T. quadrifrons from a calyx height of around 4 to around 8 mm, and in T. rarus from 17 to 22 at heights from about 8 to 11.5 mm, although the new species has more ribs than representatives of the other species at comparable sizes. We have printed a linear trendline to these epispire row-data, although the distribution of data points vaguely suggest an exponential correlation; the number of specimens and thus data is too low to draw conclusions on the exponential increase in row number. Nevertheless, a nonlinear increase in epispire rows would not be surprising since the volume of a sphere (and the calyx of Tiaracrinus is nearly spherical) also increases exponentially when compared to diameter (calyx height) increase; an exponentially growing body would probably have an exponentially growing respiration, hence the possibly exponential increase in epispire rows. The variation in strength of the ribs and size of the epispires within one specimen appears to be low (compare Hauser 2008: pl. 1).

Phylogeny

Some of the relationships within the genus are undoubtedly as proposed by Le Menn (1990). Accordingly, T. oehlerti and T. moravicus with their elongate and slightly ornamented calices represent a small monophylum (supported by the results of our cluster analysis in Fig. 5). Intrageneric variability is dominated by the significant ontogenetic increase in the number of epispire rows. A possible correlation between calyx width and height through ontogeny lacks statistical support, which is probably due to the low amount of data. Alternatively, as proposed by Hauser (2008: 2), T. oehlerti and T. moravicus could be conspecific and were assigned to different species because of differences in preservation. The same question is open with respect to T. tedraedra and T. quadrifrons, two superficially similar species with profoundly different preservation (Hauser 2008: 5), which also form one cluster. T. jeanlemenni, T. rarus, T. tedraedra, and T. quadrifrons share the broader calyx with the narrower radial channels and might thus be closely related (see Fig. 3).

Fig. 6.

Reconstruction of the phylogeny of the species of Tiaracrinus, based on their morphology and stratigraphic occurrences. Devonian timescale after Kaufmann (2006). All ranges of species lack precision and are thus indicated by open boxes.

The question arises, which of the various kinds of morphology represents the more plesiomorphic state. The age (Pragian to earliest Eifelian) of most of the reasonably well-known species does not differ significantly. Since T. oehlerti appears to be the oldest of the better known species, we suggest that the morphology with the elongate calyx and the more ornamented radial channels represent the more ancestral morphology (Fig. 6). At least the calyx shape is more similar to the genus Zophocrinus, which is one of the few genera in the same family according to the crinoid Treatise (Moore et al. 1978). Remarkably, all the other genera assigned to the Zophocrinidae lack epispires. Following Guensburg and Sprinkle (2007: 287), the presence of epispires in crinoids can be considered as a plesiomorphic character and thus, the morphology of Zophocrinus and Parazophocrinus would be more derived, if we accept the taxonomy of Moore et al. (1978). By contrast, it could be argued that the Tiaracrinus-lineage had evolved convergently, an assumption, which is less parsimonious. For a test of this latter hypothesis, material and data are not available at this point. Such material would have to show a transition between the epispire-free Zophocrinus and early Tiaracrinus with epispires, for instance with a very low number of rows.

Conclusions

Based on the higher number of ribs in the rhomboid rib-fields and the globular calyx-shape, we describe the new species Tiaracrinus jeanlemenni sp. nov., which resembles the Czech species T. rarus. The species pair T. jeanlemenni sp. nov. and T. rarus differs from the other Tiaracrinus species in the greater number of more narrow epispire rows and from the pair T. oehlerti and T. moravicus in the more globular calyx with the more narrow radial channels. Morphometric analyses (PCA and Cluster Analyses) support the closer relationships between the species pairs Tiaracrinus jeanlemenni sp. nov. and T. rarus, between T. tedraedra and T. quadrifrons, as well as between T. oehlerti and T. moravicus. In the cases of the latter two pairs, two explanations are at hand: Either, (i) T. tetraedra and T. moravicus represent junior synonyms with the type materials of both species-pairs being differently preserved or (ii) all four species, are closely related to the respective other species of the two species-pairs, representing two two-species monophyla. Based on their younger age and the much higher than wide calyx (reminiscent of the outgroup-genus Zophocrinus), we interpret the species-pair T. oehlerti and T. moravicus as the most plesiomorphic one within the genus, from which the other four species are derived.