Article: Comments on: Testing hypotheses of element loss and instability in the apparatus composition of complex conodonts (Zhang et al.)

Zhang et al. (2017) discussed the apparatus composition of the Early Triassic conodont Hindeodus parvus on the basis of six fused clusters of elements recovered from Shangsi in South China. They used the evidence from these to suggest that the previous reconstruction of the apparatus of H. parvus by Agematsu et al. (2014) is incorrect, despite being based on natural assemblages of elements from Japan. The most striking difference between the reconstructions proposed by Zhang et al. (2017, fig. 10) and Agematsu et al. (2014, fig. 6) is the re‐assignment of the P₂ elements to the S₁ position. This proposed reconstruction leaves the apparatus without P₂ elements, which is interpreted by the authors to be an example of element loss caused by changes in food capture (Zhang et al. 2017, p. 11). Agematsu et al. (2014) did not identify S₁ elements in their natural assemblages, but included them in their reconstruction; they suggested that the S₁ elements were present in the apparatus of H. parvus, but were either not visible or not preserved in their natural assemblages.

Herein, we question the new reconstruction of Zhang et al. (2017) on the basis of three main lines of evidence: (1) X‐ray CT scans of additional natural assemblages of Hindeodus parvus from Funabuseyama in Japan (Agematsu et al. 2017), which allow the recognition of 15 elements in the apparatus; (2) a previously unpublished reconstruction of the multielement apparatus of Lower Triassic Hindeodus based on discrete elements from the Lower Triassic of Heping in South China, which illustrates the morphology of both the P₂ and S₁ elements of this genus; and (3) a review of all previous reconstructions of Hindeodus, from the Mississippian to the Early Triassic, which demonstrates the striking conservatism of the apparatus composition of this genus throughout more than 100 myr. Ultimately, we find Zhang et al.'s (2017) hypothesis of element loss in H. parvus to be unsupported.

X‐ray CT scans of natural assemblages

It is commonly difficult to discriminate all of the elements in a natural assemblage. The ideal reconstruction of an apparatus is strongly dependent on the degree of fossil preservation, the orientation of compression of the rock, and the splitting direction of the rock sample. ‘Complex’ conodonts, including the Order Ozarkodinida Dzik, 1976 (sensu Donoghue et al. 2008), usually possess an S array comprising pairs of S₁, S₂, S₃, and S₄ elements that resemble each other in shape. When long processes with numerous thin denticles are clustered together in a specimen containing a natural assemblage, it is difficult to distinguish individual elements. Agematsu et al. (2014) described assemblages of Hindeodus species from the Hashikadani Formation in the Funabuseyama area, Mino Terrane, Japan, in which it was impossible to recognize S₁ elements because they were mostly buried in the claystone. Elements at the S₁ site, which is located in the innermost part of the array, were hidden by a bush of S₂, S₃, or S₄ elements, both in laterally compressed specimens (Agematsu et al. 2014, figs 4, 5) and dorso‐ventrally compressed specimens (Agematsu et al. 2014, fig. 3). However, the absence of evidence does not provide evidence of absence; consequently, we must be careful in drawing conclusions based on only a few specimens.

The Funabuseyama specimens are preserved as moulds without a body of elements. Agematsu et al. (2017) used synchrotron radiation micro‐tomography (SR–μCT) to examine the 15‐element hypothesis of Hindeodus rather than employing surface observations under a microscope. After a trial analysis of c. 20 samples they acquired well‐reconstructed three‐dimensional images of specimens EESUT‐ag0005 and ‐ag0006, which are ventral and dorsal parts, respectively, constituting a single natural assemblage. SR–μCT not only reveals elements that are hidden under the rock surface, but also the direction of element curvature and the three‐dimensional arrangement of elements. Images of specimens EESUT‐ag0005 and ‐ag0006 are largely consistent with the well‐known ozarkodinid apparatus model of Purnell & Donoghue (1998) in terms of general architecture, the number of elements, and the direction of curvature. Although both dextral and sinistral S₁ elements have unfortunately lost the distal parts of ‘outer lateral’ processes, possibly during diagenesis, they are probably digyrate elements. In any event, it is certain that the Hindeodus apparatus has the 15‐element plan.

The elements of Hindeodus parvus

In order to interpret fused clusters, it helps to be familiar with the constituent elements. It is difficult to recognize elements from two‐dimensional line‐diagrams such as those in Agematsu et al. (2014) and Zhang et al. (2017). Images of discrete elements in Zhang et al. (2017) include views of five different elements, but not of the problematic S₁/P₂ element.

Discrete elements belonging to Hindeodus present in two collections from the Lower Triassic (Griesbachian) at the Heping section, in South China (Lehrmann et al. 2003) were assembled as representative of the generic apparatus (Fig. 1). Although P₁ elements of H. typicalis are also present in both collections, and Isarcicella in one, no additional morphotypes of the ramiform elements were identified. This leads to the suggestion that the apparatus of each of these contemporaneous taxa were similar, if not identical. It follows that there is no way to determine in which species the vicarious elements originated. Nevertheless, the apparatus illustrated here is consistent with the natural assemblage from Japan as well as most previous reconstructions, including H. parvus (e.g. Kozur 1996; see below).

Figure 1 Open in figure viewerPowerPoint Reconstructed apparatus of Lower Triassic Hindeodus from the Heping section, South China. A, P₁ element of Hindeodus parvus, GSC type no. 139307. B, P₂ element, GSC type no. 139308. C, S₃/S₄ element, GSC type no. 139309. D, M element, GSC type no. 139310. E, S₁ element, GSC type no. 139311. F, S₂ element, GSC type no. 139312. G, S₀ element, GSC type no. 139313. A, C–G, come from sample HPC‐20 (@12.8 m above the base of the microbialite), which contains both H. parvus and H. typicalis; the P₂ element (B), selected for its superior preservation, comes from sample HPC‐23 (@15.3 m, immediately above the microbialite), which in addition contains Isarcicella sp. Illustrated specimens are housed at the National Type Collection of Invertebrate and Plant Fossils at the Geological Survey of Canada (GSC) in Ottawa, Ontario, Canada. Scale bar represents 200 μm. CaptionReconstructed apparatus of Lower Triassic Hindeodus from the Heping section, South China. A, P1 element of Hindeodus parvus, GSC type no. 139307. B, P2 element, GSC type no. 139308. C, S3/S4 element, GSC type no. 139309. D, M element, GSC type no. 139310. E, S1 element, GSC type no. 139311. F, S2 element, GSC type no. 139312. G, S0 element, GSC type no. 139313. A, C–G, come from sample HPC‐20 (@12.8 m above the base of the microbialite), which contains both H. parvus and H. typicalis; the P2 element (B), selected for its superior preservation, comes from sample HPC‐23 (@15.3 m, immediately above the microbialite), which in addition contains Isarcicella sp. Illustrated specimens are housed at the National Type Collection of Invertebrate and Plant Fossils at the Geological Survey of Canada (GSC) in Ottawa, Ontario, Canada. Scale bar represents 200 μm.

The angulate P₂ element is clearly present in these collections: it possesses a large cusp and an inturned posterior process and is similar to those illustrated by Agematsu et al. (2014, figs 2C, 4C, 5C) and Agematsu et al. (2017, fig. 4). The P₂ element is apparently absent from the clusters of Zhang et al. (2017). The S₁ element from the Heping collections is digyrate, with a relatively short anterior process and a longer posterior process. This element is similar to those illustrated by Agematsu et al. (2014, fig. 4D) as S₁–S₂?, and by Agematsu et al. (2017) as S₁. Furthermore, previous studies of discrete elements from the Lower Triassic of Shangsi have illustrated P₂ elements of Hindeodus parvus with a similar morphology to those from Heping (e.g. Li et al. 1989, pl. 45, fig. 3, pl. 52, figs 17, 18). This demonstrates that such elements are present in the same area as sampled by Zhang et al. (2017), even though they are absent from the fused clusters studied by those authors. The morphology of the P₂ and S₁ elements of H. parvus from South China are very similar to those of other species of Hindeodus, as discussed below.

The Hindeodus apparatus through time

The multielement apparatuses of numerous species of Hindeodus have been reconstructed from the Early Carboniferous to the Early Triassic by many different workers (Table 1). Throughout its stratigraphic range, the apparatus of this genus has shown negligible change. Although many authors have been unable to recognize all of the elements of Hindeodus, it is possible to reconstruct the complete apparatus for the genus by considering all of the published sources. Based on these reconstructions (Table 1), it can be implied that the apparatus of Hindeodus consisted of: a pair of carminiscaphate P₁ elements; a pair of angulate P₂ elements; a pair of digyrate M elements; a single alate S₀ element; a pair of digyrate S₁ elements; a pair of digyrate or bipennate S₂ elements; a pair of bipennate S₃ elements; and a pair of bipennate S₄ elements.

Table 1. Multielement apparatus reconstructions of Hindeodus from the Mississippian to the Early Triassic Element P₁ P₂ M S₀ S₁ S₂ S₃ S₄Hindeodus cristulus (Sweet & Clark 1981); Mississippian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus cristulus (von Bitter & Plint‐Geberl 1982); middle Mississippian Spathognathodiform Ozarkodiniform Neoprioniodiform Hindeodiform Not illustrated Not recognized Not illustrated Not recognized Hindeodus cristulus (Rexroad & Furnish 1964); late Mississippian Spathognathodiform (Ozarkodiniform) (Neoprioniodiform) (Hindeodiform) (Hindeodiform) Not recognized (Hindeodelliform) Not recognized Hindeodus cristulus (Rexroad & Nicoll 1965); late Mississippian Spathognathodiform (Ozarkodiniform) (Neoprioniodiform) (Hindeodiform) (Hindeodiform) Not recognized Not recognized Not recognized Hindeodus cristulus (Sweet 1977); late Mississippian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus cristulus (Rexroad 1981); late Mississippian Spathognathodiform Ozarkodiniform Neoprioniodiform Trichonodelliform Falcodiform Not recognized Hindeodelliform Not recognized Hindeodus cristulus (Horowitz & Rexroad 1982); late Mississippian Carminiscaphate Angulate Not recognized Alate Digyrate Bipennate Bipennate Not recognized Hindeodus minutus (Horowitz & Rexroad 1982); late Mississippian Carminiscaphate Angulate Not recognized Alate Digyrate Bipennate Bipennate Not recognized Hindeodus minutus (Merrill & Powell 1980); early Pennsylvanian Carminiscaphate Ozarkodinid Neoprioniodinid Not recognized Plectospathodid Not recognized Hindeodellid Not recognized Hindeodus minutus (Stamm & Wardlaw 2003): middle Pennsylvanian Carminiscaphate Angulate Dolobrate Alate Digyrate Bipennate Bipennate Not recognized Hindeods minutus (von Bitter 1972); late Pennsylvanian Anchignathodontan Ozarkodinian (Dolobrate) Hindeodontan (Digyrate) Not recognized (Bipennate) Not recognized Hindeodus minutus (Baesemann 1973); late Pennsylvanian Spathognathodontan Ozarkodinan Neoprioniodontan Hindeodontan (Digyrate) Not recognized Hindeodellan (Bipennate) Hindeodus minutus (Perlmutter 1975); late Pennsylvanian Ozarkodinan (Angulate) (Neoprioniodontan) Hindeodontan Not recognized Not recognized (Bipennate) Not recognized Hindeodus minutus (Merrill 1980); late Pennsylvanian Spathognathodiform Ozarkodiniform Neoprioniodiform Trichonodelliform Plectospathodiform Not recognized Hindeodelliform Not recognized Hindeodus minutus (Rasmussen & Håkansson 1996); early Permian Carminiscaphate Angulate Not illustrated Not illustrated Not illustrated Not illustrated Not illustrated Not illustrated Hindeodus sp. (Clark & Carr 1982); early Permian Spathognathodiform Ozarkodiniform Neoprioniodiniform Hindeodiform (Digyrate) Not recognized Hindeodelliform Not recognized Hindeodus sp. (von Bitter & Merrill 1985); Carboniferous–Permian Carminiscaphate Angulate Dolobrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus typicalis (Babcock 1976); middle Permian Carminiscaphate (Angulate) (Dolobrate) (Alate) (Digyrate) Not recognized (Bipennate) Not recognized Hindeodus excavatus (Sweet 1977); middle Permian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus excavatus (Wardlaw & Collinson 1984); middle Permian Carminiscaphate Angulate Dolobrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus excavatus (Wardlaw 2000); middle Permian Carminiscaphate Angulate Dolobrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus typicalis (Sweet 1970); late Permian Carminiscaphate Angulate Dolobrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus julfensis (Sweet 1977); late Permian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus julfensis (Nestell & Wardlaw 1987); late Permian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus julfensis (Metcalfe 2012); late Permian Carminiscaphate Angulate Digyrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus latidentatus (Wardlaw et al. 2015); late Permian Carminiscaphate Angulate Dolobrate Alate Digyrate Bipennate Bipennate Bipennate Hindeodus sp. (Kozur 1975); late Permian – Early Triassic Anchignathodiform Not recognized Neoprioniodiform Not recognized Enantiognathiform Not recognized Hindeodelliform Not recognized Hindeodus latidentatus (Kozur et al. 1995); Early Triassic Carminiscaphate Not illustrated Not illustrated Not illustrated Not illustrated Digyrate Not illustrated Not illustrated Hindeodus parvus (Kozur & Pjatakova in Kozur 1977); Early Triassic Anchignathodiform Not recognized Neoprioniodiform Not recognized Not recognized Not recognized Hindeodelliform Not recognized Hindeodus minutus (Matsuda 1981); Early Triassic Carminiscaphate Angulate Dolobrate Alate Digyrate Not recognized Bipennate Not recognized Hindeodus parvus (Kozur 1995); Early Triassic Carminiscaphate Angulate Dolobrate Digyrate Not illustrated Bipennate Not illustrated Not illustrated Hindeodus parvus (Kozur 1996); Early Triassic Carminiscaphate Angulate Dolobrate Alate Digyrate Not illustrated Bipennate Not illustrated Hindeodus parvus (Perri et al. 2004); Early Triassic Carminiscaphate Not illustrated Not illustrated Not illustrated Digyrate Not recognized Bipennate Not recognized Hindeodus parvus (Yang et al. 2014); Early Triassic Carminiscaphate Not illustrated Dolobrate Alate Not illustrated Bipennate Bipennate Bipennate Hindeodus parvus (Agematsu et al. 2014); Early Triassic Carminiscaphate Angulate Makellate Alate Not recognized Digyrate Bipennate Bipennate Hindeodus parvus (Zhang et al. 2017); Early Triassic Carminiscaphate Not recognized Makellate Alate Digyrate Digyrate Bipennate Bipennate

• Anatomical notation for the elements follows Purnell et al. (2000). Not recognized = authors do not assign any elements to this anatomical position within the apparatus reconstruction. Not illustrated = authors indicate that an element belongs in this anatomical position, but do not illustrate or describe it. The morphological notation of the original authors is retained, except for when no written description of the elements is provided, in which case the morphological terms of Sweet (1988) are applied. Those elements in parentheses are illustrated by the authors, but not assigned to a particular anatomical position within the Hindeodus apparatus.

As mentioned previously, disagreement between the reconstructions of Agematsu et al. (2014) and Zhang et al. (2017) primarily concerns the anatomical placement of the angulate elements in the natural assemblages from Japan, interpreted as P₂ elements by Agematsu et al. (2014). The P₂ element of Hindeodus has previously been demonstrated to be an angulate element in the middle to late Mississippian H. cristulus (Sweet 1977; Horowitz & Rexroad 1982; von Bitter & Plint‐Geberl 1982); in the late Mississippian to early Permian H. minutus (von Bitter 1972; Baesemann 1973; Merrill 1980; Horowitz & Rexroad 1982; Rasmussen & Håkansson 1996); in the middle Permian H. excavatus (Sweet 1977; Wardlaw & Collinson, 1984; Wardlaw 2000); in the late Permian H. julfensis (Sweet 1977; Nestell & Wardlaw 1987; Metcalfe 2012) and H. latidentatus (Wardlaw et al. 2015); and it has even previously been recognized in H. parvus from the Early Triassic (Kozur 1996). This strongly suggests that the angulate elements identified in the bedding plane assemblages of Hindeodus parvus by Agematsu et al. (2014) are P₂ elements.

Phylogenetic position of ‘Hindeodus’ parvus

If Hindeodus parvus did indeed lack P₂ elements, then its position within the genus Hindeodus is questionable. It has been demonstrated that the type species of the genus, H. cristulus, possessed P₂ elements (see above), and therefore parvus should be excluded from Hindeodus. This contradicts a previous cladistic study by two of the authors of the Zhang et al. (2017) paper based on the morphology of P₁ elements (Jiang et al. 2011), which places parvus within Hindeodus, most closely related to H. anterodentatus, H. sosioensis and H. postparvus. This group is in turn interpreted to be phylogenetically close to Isarcicella by Jiang et al. (2011). The multielement apparatuses of these species of Hindeodus have not been reconstructed, and therefore it is plausible that they too could lack P₂ elements. The apparatus of Isarcicella has also not been reconstructed, although as noted above it is probably very similar, if not identical, to that of Hindeodus; Yang et al. (2014) suggest that it is similar to that of Parafurnishius, which possesses P₂ elements. The phylogenetic position of H. parvus as determined by Jiang et al. (2011) implies that it possessed P₂ elements as well.

Summary

The apparatus reconstructions of Hindeodus parvus by Agematsu et al. (2014, 2017) are favoured over that of Zhang et al. (2017). The former reconstructions are consistent with evidence from natural assemblages and collections of discrete elements belonging to H. parvus, and conform to the apparatus pattern displayed by all other members of Hindeodus. These reconstructions also support the previously determined phylogenetic position of H. parvus as a derived species within the genus. The failure to recover P₂ elements from one collection at Shangsi does not prove the absence of this element from the apparatus, when it has been recognized in numerous other collections of Hindeodus. The hypothesis of element loss from the P domain of this genus is therefore unsupported by the majority of available data.

Institutional abbreviations

EESUT, Earth Evolution Sciences, University of Tsukuba, Japan.

Acknowledgements

D. J. Lehrmann (Trinity University, Texas) collected the samples from the Heping section, and is thanked for allowing this material to be published. We would also like to thank the editor and an anonymous reviewer for comments that helped to improve this manuscript. Partial funding for this work came from the Geological Survey of Canada GEM 2 program.